The influence mechanisms of living environment on the physical and mental health of Chinese older adults: a cross-sectional exploration based on CLASS 2023 Authors: Ye Wu Published: 2025 ABSTRACT Against the backdrop of continuously deepening population aging, the physical and mental health of older adults has become a significant research topic in academic circles. However, existing studies have not yet established a clear understanding of the mechanisms linking living environments to the physical and mental health of older adults, and a notable gap remains regarding gender-specific variations in these associations. Based on the social ecology theoretical framework, this study uses 9998 valid samples from the China Longitudinal Aging Social Survey (CLASS 2023) to construct a theoretical analysis model of “living environment—social adaptation—physical and mental health,” and systematically explores the complex relationships among living environments, social adaptation, and the physical and mental health of older adults through structural equation modeling. The findings reveal that living environment is significantly positively correlated with the physical and mental health of older adults, with a notably stronger correlation to mental health than to physical health. Social adaptation plays a partial mediating role in the process through which the built environment affects physical and mental health. Gender difference analysis reveals that the correlation between the living environment and the physical and mental health as well as social adaptation is more pronounced among male older adults.The research conclusions provide empirical evidence for differentiated aging-friendly environmental planning, and it is recommended that gender-differentiated environmental intervention strategies be established in urban planning practices to promote the overall health of older adults.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-26254-z. Introduction Against the backdrop of the accelerating global trend of population aging, the maintenance of physical and mental health among older adults has become a core research topic1. As the country with the largest older adult population in the world, China faces severe aging challenges: by the end of 2023, the population aged 60 and above reached 297 million, accounting for 21.1% of the total population, marking the arrival of a deeply aging society2. The structural contradictions in the health dimension brought by an aging society are particularly prominent: the population of disabled older adults exceeded 35 million in 2024, accounting for 11.8% of the older adult population, significantly higher than the global average disability rate; mental health surveys show that 23.76% of Chinese older adults exhibit depressive symptoms of varying degrees, and female older adults demonstrate higher vulnerability in physiological decline and psychological issues3. Against this backdrop, China’s Healthy China 2030 plan explicitly proposes the strategic goal of “healthy aging,” incorporating the optimization of aging-friendly built environments into the core agenda of urban planning, thus providing a new spatial governance pathway for addressing challenges in older adult health4. Against the accelerating backdrop of urbanization and rural revitalization strategies, the profound transformation of China’s living environment is significantly reshaping residents’ lifestyles and health outcomes. Extensive empirical research has confirmed that optimizing the living environment is a crucial pathway to enhancing both physical and mental health. Scholars have explored the relationship between the living environment and health from various dimensions, including green landscapes, road traffic, and infrastructure availability. For example, Tabrizi et al.5 found that green spaces have a stress-buffering effect, significantly reducing the negative impacts of harmful environmental exposures on physical health and improving mental health; Li et al.’s6 empirical study showed that high-quality nighttime walking environments can reduce the prevalence of chronic diseases by promoting physical activity; Sundling and Jakobsson’s7 meta-analysis further revealed a significant negative correlation between the quality of walking environments and negative emotions such as depression and anxiety. The spatial accessibility and equitable allocation of convenient facilities directly influence their health and well-being8,9. Nevertheless, a persistent research paradox remains under-addressed: although older adults constitute one of the most extensive and frequent user groups of urban and rural spaces, they are simultaneously among the most vulnerable to environmental influences on health. Current research exhibits two major limitations. First, there is a notable scarcity of systematic investigations targeting the national aging population. Second, extant studies tend to focus on isolated health dimensions—such as physical or mental health examined separately—rather than adopting an integrated perspective that accounts for their interplay within a socio-ecological framework. This fragmented approach falls short of elucidating the systemic mechanisms through which living environments impact older adults’ overall well-being via behavioral mediators. Consequently, these gaps represent critical blind spots in both theoretical knowledge and practical intervention strategies. As research deepens, there is increasing academic recognition that the relationship between the living environment and population health is far from a simple bivariate correlation, but rather involves complex mediating mechanisms. Social Ecological Theory provides a core framework for parsing these mechanisms, emphasizing a hierarchical pathway through which macro-environmental systems influence micro-level health outcomes by shaping individual behavioral patterns10. From this theoretical perspective, environmental behavior theory further clarifies that the built environment guides or constrains individuals’ daily activity choices—such as social participation and environmental adaptation—by providing specific “behavioral settings” and “opportunity structures”11. Thus, the health status of older adults is not only a direct product of their living environment but, more importantly, reflects a mediation effect whereby the environment influences health through shaping behavioral patterns—particularly social and physical behaviors12. Extensive research has confirmed that well-designed community environments effectively promote moderate exercise among older adult, thereby significantly enhancing cardiopulmonary function, quality of life, and cognitive performance13. However, compared with physical behaviors, social behaviors serve as critical ties for maintaining social networks, acquiring emotional support, and fulfilling self-realization in older adults, making them a more central element within their health ecosystem14. Unfortunately, existing studies have rarely incorporated social behavior systematically as a core mediating variable into theoretical models exploring the impact of the living environment on health of older populations. Therefore, it is imperative that future research prioritizes and empirically examines the key mediating role of social behavior within the complex mechanism through which living environments (macro-system) influence the health (micro-level outcomes) of older adults, so as to address major gaps in both theoretical understanding and practical intervention. As a key social stratification variable, gender exhibits significant heterogeneity in the field of older adult health. Within the specific socio-cultural context of China, health disparities among older adults vividly reflect the influence of social structure and gender norms: female older adults have longer average life expectancies but higher postmenopausal prevalence rates of chronic diseases (such as arthritis and osteoporosis), and are more prone to depressive tendencies due to social role transitions (such as empty nesting and spousal loss); males face higher risks of cardiovascular diseases and sarcopenia, with mental health issues often masked by traditional gender norms15,16. The impact of the living environment on older adults’ health also demonstrates marked gender differentiation. Studies indicate that green spaces, walkable areas, and social facilities contribute to health in distinct ways across gender groups17. Nevertheless, there remains a dual mechanistic blind spot in current research: first, whether social adaptation plays differential mediating roles in the environment-health relationship across gender strata; second, what differences exist in the intensity of direct and indirect impacts of built environments across gender groups. Existing research often overlooks these underlying gender-specific mechanisms, resulting in “gender-neutral” planning interventions that fail to adequately address differentiated needs and may even perpetuate health inequalities. This study seeks to deconstruct the gendered patterns in environmental exposure, behavioral response, and health outcomes among older adults in China, thereby providing a scientific basis for equity-oriented living environment design. This study constructs a conceptual model of “living environment—social adaptation—physical and mental health” using structural equation modeling (SEM) and data from the China Longitudinal Aging Social Survey (CLASS) 2023. It examines the associative relationships among living environments, social adaptation, and physical and mental health in older adults, as well as variations across genders. The results indicate that more favorable living environments are associated with higher levels of social adaptation and better health outcomes in older adults, thereby extending the discourse on health-oriented human settlements planning and offering insights relevant to rural revitalization and ecological civilization initiatives. Furthermore, the findings reveal gender-based differences in how living environments correlate with social adaptation and health outcomes, helping to clarify gendered behavioral patterns and providing empirical support for the development of age-friendly living environments. Methods Data and samples The data for this study were derived from the China Longitudinal Aging Social Survey (CLASS) hosted by Renmin University of China. The survey adopted a multistage stratified probability sampling design, covering 28 provinces/autonomous regions/municipalities directly under the Central Government nationwide, aiming to obtain multidimensional data on older adults’ health, environments, and social adaptation through continuous tracking. This study utilized the 2023 tracking data, with the sampling process as follows: first, villages/communities were used as primary sampling units, and sample points were randomly selected based on a map address framework; second, households were randomly selected within the sample points; finally, the Kish table method was employed to sample individuals aged 65 and above within households, resulting in 11,670 face-to-face interviews. During the data preprocessing stage, missing values were addressed using listwise deletion. We began by diagnosing missing values across all core variables—both dependent and independent—as well as covariates included in the analyses. Given the low proportion of missing data among the study variables, and since Little’s MCAR test results suggested that the missing data pattern conformed to the missing completely at random (MCAR) assumption, the application of listwise deletion was considered both appropriate and methodologically sound. This strategy ensured that the final analysis was conducted on a complete and consistent dataset, thereby circumventing potential uncertainties associated with multiple imputation. In total, 1,672 cases with missing key variables were excluded, yielding a final analytical sample of 9,998 cases and a sample retention rate of 87.4%. The sample structure featured: 5,198 males (52.0%), 4,800 females (48.0%) and educational attainment predominantly at primary school level. This sample is nationally representative and meets the sample selection criteria for older adult research. The CLASS is a sociological research project and does not fall within the realm of medical research. It adheres to the ethical standards outlined in the Declaration of Helsinki and does not involve human experimentation as defined by that Declaration. Furthermore, the questionnaire is specifically designed to avoid any adverse effects on the health of the respondents. Consequently, the CLASS survey does not require additional ethical approval. All participants provided written informed consent, and the information collected is intended exclusively for sociological research purposes, with strict measures in place to ensure confidentiality and anonymity. The detailed descriptions of study design and implement have been published previously, and are available at the study website (http://class.ruc.edu.cn/English/Home.htm). Variables and measurements Dependent variable 1: physical health The first dependent variable in this study was physical health, measured using the Instrumental Activities of Daily Living (IADL) scale recommended by the World Health Organization18–20. The IADL scale assesses physical health through eight daily living abilities: climbing stairs, walking outdoors, using transportation, shopping outside, managing finances, lifting heavy objects, doing housework, and cooking. The scale uses a 3-point scoring system (1 = fully independent, 2 = partially dependent, 3 = fully dependent), with a total score range of 8–24 points; higher scores indicate more severe disability. Reliability and validity tests showed a Cronbach’s α coefficient of 0.901 and a Kaiser-Meyer-Olkin (KMO) value of 0.885 (p < 0.001), indicating good internal consistency and construct validity. Dependent variable 2: mental health The other dependent variable was mental health, measured using the Center for Epidemiologic Studies Depression Scale (CES-D) revised by Cong and Silverstein21, which has been widely validated in mental health research and clinical applications22–24. Through the CES-D scale, older adults were asked to report the frequency of six emotions over the past week: (1) feeling depressed, (2) feeling lonely, (3) feeling useless, (4) poor sleep, (5) poor appetite, and (6) feeling unhappy with life. Each item has three response options: 1 = never, 2 = sometimes, 3 = often; higher scores indicate more severe depressive symptoms. Reliability and validity analysis of the mental health scale included in the study showed a Cronbach’s Alpha of 0.802 and a KMO of 0.868 (p < 0.001), indicating strong reliability and validity for measuring mental health in older adults using CES-D. Independent variable: living environment In the CLASS 2023 survey, living environments were assessed through subjective evaluations, which have become a primary evaluation method in architectural planning in recent years25,26. Obtaining environmental data via subjective evaluation not only directly reflects individuals’ real experiences and emotional perceptions of their environments but also efficiently captures diverse needs. Groups differing in gender, age, occupation, and health status exhibit significant stratification in environmental perceptions, and subjective evaluation can accurately represent this diversity, providing a basis for refined policy-making and enabling data to comprehensively and three-dimensionally reflect real environmental landscapes. Living environment in this study was measured across six dimensions: road traffic, fitness/activity venues, public security environment, environmental sanitation, road/street lighting, and barrier-free facilities. Responses used a 5-point scale from “very satisfied” to “very dissatisfied,” scored 1–5 points respectively; lower scores indicate higher environmental quality. Reliability and validity analysis of the included scale showed a Cronbach’s Alpha of 0.818 and a KMO of 0.869 (p < 0.001), indicating strong reliability and validity for measuring living environments. Mediating variables: social adaptation The CLASS 2023 database used in this study measures older adults’ social adaptation levels through the Social Adaptation Scale developed by Bosc27–29. The scale includes six items: “I feel I am useless to society,” “Social changes are too fast for me to adapt,” “More and more viewpoints nowadays are difficult for me to accept,” “More and more new social policies today are difficult for me to accept,” “Current social changes are increasingly unfavorable to older adults,” and “I can no longer keep up with the pace of the times.” The scale uses a 5-point scoring system (1 = completely inconsistent, 5 = completely consistent), where higher scores indicate lower levels of social adaptation among older adults. In this study, the scale had a Cronbach’s α coefficient of 0.865, and a KMO of 0.860 (p < 0.001), demonstrating reliable measurement stability. Control variables In examining the impact pathway of living environments on older adults’ physical and mental health, the model may be influenced by confounding factors. To enhance the credibility and robustness of the results, this study included educational attainment, gender, age, ethnicity, economic status, number of co-residents, self-rated health, health change, and body mass index (BMI) as control variables to reduce the influence of confounding factors on research outcomes30,31. Specifically: educational attainment was scored on a scale of 1–7, ranging from illiterate to above bachelor’s degree. Gender was coded as 1 = male, 2 = female. Ethnicity was coded as 1 = Han, 2 = Mongolian, 3 = Manchu, 4 = Hui, 5 = Tibetan, 6 = Zhuang, 7 = Uyghur, 8 = other. Economic status was rated as 1 = good, 2 = average, 3 = poor. Number of cohabitants ranged from 1 to 10. Self-rated health among older adults was assessed with: 1 = very healthy, 2 = relatively healthy, 3 = average, 4 = relatively unhealthy, 5 = very unhealthy. Health change was evaluated as: 1 = improved, 2 = unchanged, 3 = worsened. BMI was calculated as weight in kilograms divided by height in meters squared (kg/m²). Research analysis This study employs SEM to explore the complex relationships among living environments, social adaptation, and older adults’ physical and mental health, offering multiple advantages: First, it effectively addresses latent variable measurement issues. Living environments, social adaptation, and physical/mental health are all multidimensional abstract concepts. SEM integrates multi-dimensional observed indicators through confirmatory factor analysis, accurately depicting the essential characteristics of latent variables while separating measurement errors to enhance the reliability and validity of research constructs32. Second, it systematically examines the associative relationships among variables, enabling the estimation of both the direct associations between living environments and physical and mental health and the mediating role of social adaptation, as well as the identification of potential moderating roles. This approach provides a comprehensive understanding of the direct, indirect, and overall linkages among these factors33. Third, it possesses robust model fitting and comparison capabilities, evaluating the goodness-of-fit between theoretical models and data through fit indices and supporting nested model comparisons to facilitate selection of the optimal explanatory model and avoid the one-sidedness of simple linear regression34. The study further conducted multi-factor confirmatory analyses on the measurement models for living environments, social adaptation, and older adults’ physical and mental health. The composite reliabilities of the four measurement models were 0.819, 0.886, 0.924, and 0.791, all meeting the standard of > 0.700; the standardized factor loadings and reliability coefficients (SMC) of observed variables all exceeded or approached the standards of 0.600 and 0.36035. Therefore, all measurement models exhibited good reliability and validity, making them suitable for structural equation modeling analysis. This study employed structural equation modeling (SEM) using Mplus 8.0, with the weighted least squares mean and variance adjusted (WLSMV) estimator applied to handle ordered categorical variables. Confidence intervals were computed via Bootstrap sampling with 1000 iterations. Table 1 presents the model fitting results, showing that the fit indices RMSER < 0.080 (Root mean square error of approximation), SRMR < 0.080 (Standardized root mean square residual), CFI > 0.90 (Comparative fit index), and TIF > 0.900 (Tucker-Lewis index) all Meet ideal standards, indicating excellent model fit36. Table 1Comparison of model fit metrics.TFICFIRMSEASRMR Model data 0.912 0.910 0.045 0.033 Idealized standard > 0.900 > 0.900 < 0.080 < 0.080 Table 1Comparison of model fit metrics.TFICFIRMSEASRMR Model data 0.912 0.910 0.045 0.033 Idealized standard > 0.900 > 0.900 < 0.080 < 0.080 Comparison of model fit metrics. Results Descriptive statistics of research variables According to the descriptive statistical results of the research variables (Table 2), all observed variables in the living environment scored around 2 points, indicating relatively high local living environment quality but with room for improvement. Among them, barrier-free facilities and fitness/activity venues had lower scores, reflecting the need to strengthen the construction of aging-friendly hardware facilities. In terms of older adults’ physical health, all observed variables scored above 1 point, indicating existing functional issues, with particularly weaker abilities in lifting heavy objects, using transportation, and shopping. This suggests that older adults exhibit weaker performance in complex daily activities, and the mean values of all indicators for female older adults were significantly higher than those for males, indicating more pronounced physical function decline in females. For mental health, all observed variables averaged around 2 points, suggesting a tendency toward obvious psychological issues among older adults, with female older adults showing relatively worse mental states than males. Observed variables for social adaptation all scored above 3 points, indicating weak adaptability to social changes among older adults, especially females. Overall, the findings indicate that female older adults are more vulnerable in terms of physical and mental health as well as social adaptation. In terms of control variables, the study shows a nearly equal gender ratio among older adults, with an average age between 73 and 74 years. Educational attainment is predominantly primary school, with males having higher educational levels than females, and the vast majority being of Han ethnicity. Older adults’ economic status follows a normal distribution, with males generally in slightly better financial conditions. The average number of co-residents with older adults is 2–3 persons, with female older adults having more co-residents than males. Older adults’ self-rated health averaged between “relatively healthy” and “average,” with males reporting better self-rated health overall. In terms of perceived health changes, a general decline was reported among the older adult population, with a more pronounced decline observed among female respondents. The average BMI was 23.096, indicating that most older adults in the sample fall within the normal range, with women showing slightly better results on average. These descriptive results are consistent with China’s population census data, confirming the credibility and scientific validity of the research data37. Table 2Variable descriptive statistics.Observed variablesVariable itemsMean scoresMean (All)Mean (Male)Mean (Female)Physical HealthPH1Are you able to go up and down stairs?1.0431.0391.047PH2Are you able to walk outdoors?1.0341.0321.036PH3Are you able to use transportation independently?1.1031.0951.111PH4Are you able to go shopping independently?1.0801.0701.090PH5Are you able to manage your finances independently?1.0381.0351.041PH6Are you able to lift objects weighing 5 kg (approx. 10 pounds)?1.1351.1101.163PH7Are you able to prepare meals independently?1.0561.0571.055PH8Are you able to perform housework independently?1.0591.0581.060PH-sumPhysical health total score.8.5488.4968.603Mental HealthMH1How have you felt emotionally over the past week?1.7451.7331.758MH2Have you felt lonely over the past week?1.8161.7921.842MH3Do you feel unable to contribute or that you are no longer useful?1.9401.9181.963MH4How have you felt about your life over the past week?1.9331.9241.941MH5Have you experienced a lack of appetite over the past week?1.8161.8041.830MH6How has your sleep been over the past week?1.9281.9041.954Living EnvironmentLE1Road traffic conditions in the community.1.9631.9601.967LE2Fitness/activity facilities in the community.2.1142.1082.119LE3Public security/safety in the community.2.0342.0442.022LE4Environmental hygiene/sanitation in the community.2.1042.1062.102LE5Road/street lighting in the community.2.1622.1642.159LE6Barrier-free facilities in the community.2.3432.3432.343Social AdaptationSA1I feel that I am no longer useful to society.3.2643.2383.288SA2Society is changing too rapidly; I find it hard to adapt to this overwhelming pace.3.2843.3013.267SA3Nowadays, I find it increasingly difficult to accept the multitude of prevailing viewpoints.3.2403.2313.250SA4I find it increasingly difficult to accept the new social policies emerging today.3.1803.1733.188SA5Current societal changes are becoming increasingly disadvantageous for the elderly.3.1863.1773.195SA6I feel I am falling behind the times / unable to keep pace with the times.3.4073.4043.410Control VariablesGenderYour gender is?1.488––AgeYour age is?73.32173.33773.304Educational AttainmentYour educational attainment is?3.2203.3573.073EthnicityYour ethnicity is?1.3461.3521.339Number of co-residentHow many people live and eat with you?2.5572.5442.571Economic StatusHow do you rate your own economic status?1.9591.9551.965Self-rated healthHow would you describe your current health status?2.4322.3992.467Health changeCompared to last year, what changes do you feel have occurred in your health?2.1212.1092.134BMIBMI = weight (kg) / height (m)223.09623.20822.975 Table 2Variable descriptive statistics.Observed variablesVariable itemsMean scoresMean (All)Mean (Male)Mean (Female)Physical HealthPH1Are you able to go up and down stairs?1.0431.0391.047PH2Are you able to walk outdoors?1.0341.0321.036PH3Are you able to use transportation independently?1.1031.0951.111PH4Are you able to go shopping independently?1.0801.0701.090PH5Are you able to manage your finances independently?1.0381.0351.041PH6Are you able to lift objects weighing 5 kg (approx. 10 pounds)?1.1351.1101.163PH7Are you able to prepare meals independently?1.0561.0571.055PH8Are you able to perform housework independently?1.0591.0581.060PH-sumPhysical health total score.8.5488.4968.603Mental HealthMH1How have you felt emotionally over the past week?1.7451.7331.758MH2Have you felt lonely over the past week?1.8161.7921.842MH3Do you feel unable to contribute or that you are no longer useful?1.9401.9181.963MH4How have you felt about your life over the past week?1.9331.9241.941MH5Have you experienced a lack of appetite over the past week?1.8161.8041.830MH6How has your sleep been over the past week?1.9281.9041.954Living EnvironmentLE1Road traffic conditions in the community.1.9631.9601.967LE2Fitness/activity facilities in the community.2.1142.1082.119LE3Public security/safety in the community.2.0342.0442.022LE4Environmental hygiene/sanitation in the community.2.1042.1062.102LE5Road/street lighting in the community.2.1622.1642.159LE6Barrier-free facilities in the community.2.3432.3432.343Social AdaptationSA1I feel that I am no longer useful to society.3.2643.2383.288SA2Society is changing too rapidly; I find it hard to adapt to this overwhelming pace.3.2843.3013.267SA3Nowadays, I find it increasingly difficult to accept the multitude of prevailing viewpoints.3.2403.2313.250SA4I find it increasingly difficult to accept the new social policies emerging today.3.1803.1733.188SA5Current societal changes are becoming increasingly disadvantageous for the elderly.3.1863.1773.195SA6I feel I am falling behind the times / unable to keep pace with the times.3.4073.4043.410Control VariablesGenderYour gender is?1.488––AgeYour age is?73.32173.33773.304Educational AttainmentYour educational attainment is?3.2203.3573.073EthnicityYour ethnicity is?1.3461.3521.339Number of co-residentHow many people live and eat with you?2.5572.5442.571Economic StatusHow do you rate your own economic status?1.9591.9551.965Self-rated healthHow would you describe your current health status?2.4322.3992.467Health changeCompared to last year, what changes do you feel have occurred in your health?2.1212.1092.134BMIBMI = weight (kg) / height (m)223.09623.20822.975 Variable descriptive statistics. Analysis of the fitting results of the full sample model When a mediating variable exists in the model, the relationship between the independent variable and the dependent variable should be expressed as total effect, direct effect, and indirect effect. Table 3, STable1, and Fig. 1 present the results of the full example model, including effect values, confidence intervals, p-values, and other relevant statistics. After controlling for demographic and socioeconomic variables, the total, direct, and indirect associations between the living environment and physical health (0.056, 0.052, 0.004) as well as mental health (0.187, 0.172, 0.015) of older adults were all significant, with a stronger association observed for mental health. This suggests that the living environment is not only directly associated with the physical and mental health of older adults but may also be linked to health outcomes through the pathway of social adaptation. These results indicate that social adaptation serves as a partial mediator in the relationship between the living environment and both physical and mental health among older adults. Additionally, a significant direct association was observed between the living environment and social adaptation (effect size = 0.045). Social adaptation was also significantly directly associated with physical health (effect size = 0.090) and mental health (effect size = 0.340). Among the control variables, age showed significant positive associations with both physical health and mental health, with correlation coefficients of 0.200 and 0.066, respectively. This suggests that within the sample of older adults, older age is associated with higher health risks, and increasing age is correlated with declines in physical function and potential increases in psychological stress. The number of co-residents showed statistically significant positive correlations with both the physical and mental health of older adults, with correlation coefficients of 0.083 and 0.045, respectively. However, it is plausible that these associations may reflect underlying adverse factors in larger or multigenerational households—such as caregiver burden and stress arising from crowded living conditions—rather than purely beneficial effects. Specifically, although more family members may provide certain forms of life support, they could also intensify caregiving responsibilities for older adults and lead to persistent psychological stress due to limited living space and reduced privacy. Thus, more complex family structures and intergenerational relationships may exert latent adverse effects on older adults’ physical and mental well-being, potentially through mechanisms such as family conflict, caregiving pressure, and constraints on personal space38. Self-rated health was significantly correlated with both physical and mental health, with coefficients of 0.162 and 0.093, respectively. This suggests that older adults’ perception of their health is associated with their objective health outcomes. Perceived change in health also showed significant correlations with physical and mental health, with values of 0.114 and 0.076, respectively. This indicates that older adults’ assessment of their health trajectory may serve as a meaningful indicator that corresponds with their actual health status. BMI was significantly positively associated with physical health, with a correlation coefficient of 0.032. This implies that higher BMI may be related to increased risk of physical illness or reflect greater physical burden. Fig. 1Overall sample model standardisation path diagram. Overall sample model standardisation path diagram. Table 3Total, direct and indirect effects of the full sample model.Model pathEffect typeEffect estimate95% CIP-valueMediation test resultLowerUpperLiving environment-social adaptation-physical healthTotal Effect0.0560.0310.0760.000Partial mediating effectDirect Effect0.0520.0290.0730.000Indirect Effect0.0040.0020.0100.018Living environment-social adaptation-mental healthTotal Effect0.1870.1520.2150.000Partial mediating effectDirect Effect0.1720.1440.2000.000Indirect Effect0.0150.0080.0250.000Living environment-social adaptationDirect Effect0.0450.0230.0610.003–Social adaptation-physical healthDirect Effect0.0900.0640.1170.000–Social adaptation-mental healthDirect Effect0.3400.2970.3820.000– Table 3Total, direct and indirect effects of the full sample model.Model pathEffect typeEffect estimate95% CIP-valueMediation test resultLowerUpperLiving environment-social adaptation-physical healthTotal Effect0.0560.0310.0760.000Partial mediating effectDirect Effect0.0520.0290.0730.000Indirect Effect0.0040.0020.0100.018Living environment-social adaptation-mental healthTotal Effect0.1870.1520.2150.000Partial mediating effectDirect Effect0.1720.1440.2000.000Indirect Effect0.0150.0080.0250.000Living environment-social adaptationDirect Effect0.0450.0230.0610.003–Social adaptation-physical healthDirect Effect0.0900.0640.1170.000–Social adaptation-mental healthDirect Effect0.3400.2970.3820.000– Total, direct and indirect effects of the full sample model. Comparison of model paths of different gender groups Given the significant differences in health characteristics of older adults of different genders, this study used a Multi-group SEM to test the gender-differentiated characteristics of the “living environment - social adaptation - physical and mental health” pathway. By setting males (Group 1) and females (Group 2) as comparison groups, nested model tests revealed significant differences in fit between the unconstrained model and the structural coefficient equivalence model, indicating differences in path coefficients across gender groups. Table 4, STable2, Fig. 2a, and b compare the model fit results based on samples of older adults of different genders. For male older adults, the total, direct, and indirect associations between the living environment and physical health (0.068, 0.063, 0.005) and mental health (0.200, 0.182, 0.018) were all significant. This pattern suggests that the living environment is associated with improvements in physical and mental health partially through social adaptation among male older adults, with a relatively stronger direct association between the environment and health outcomes. Thus, social adaptation appears to function as an important partial mediating variable in the relationship between the living environment and health outcomes for males. Similarly, for female older adults, the total, direct, and indirect associations between the living environment and mental health (0.174, 0.160, 0.014) were all significant. This indicates that the living environment is associated with improvements in mental health partly through enhanced social adaptation, though a stronger direct association is observed between the living environment and mental health itself. These findings suggest that social adaptation also serves as a partial mediator in the pathway linking the living environment to mental health among older women. Regarding physical health, a significant direct association was found between the living environment and the physical health of female older adults (0.042), with no notable mediating effect through social adaptation. Furthermore, in the “living environment—social adaptation” relationship, the association was stronger among older males than females, with values of 0.056 and 0.040, respectively. In the relationship between social adaptation and physical health, social adaptation showed a stronger correlation with physical health among older men, with values of 0.088 and 0.043 for males and females, respectively. Conversely, social adaptation demonstrated a stronger association with mental health among female older adults compared to male, with values of 0.349 and 0.330, respectively. Fig. 2(a) Path standardisation coefficients for model paths for male older adults. (b) Path standardisation coefficients for female older adults. (a) Path standardisation coefficients for model paths for male older adults. (b) Path standardisation coefficients for female older adults. Table 4Comparison of the model paths in different gender groups.Model pathEffect typeEffect estimate95% CIP-valueMediation test resultLowerUpperMale group Living environment-social adaptation-physical healthTotal effect0.0680.0380.0870.000Partial mediating effectDirect effect0.0630.0300.0920.000Indirect effect0.0050.0010.0110.038 Living environment-social adaptation-mental healthTotal effect0.2000.1540.2360.000Partial mediating effectDirect effect0.1820.1430.2250.000Indirect effect0.0180.0090.0340.008 Living environment-social adaptationDirect effect0.0560.0230.0790.005– Social adaptation-physical healthDirect effect0.0880.0480.1270.000– Social adaptation-mental healthDirect effect0.3300.2660.3830.000–Female group Living environment-social adaptation-physical healthTotal effect0.0420.0150.0720.007No mediating effectDirect effect0.0410.0130.0710.009Indirect effect0.001− 0.0010.0030.281 Living environment-social adaptation-mental healthTotal effect0.1740.1270.2110.000Partial mediating effectDirect effect0.1600.1190.2010.000Indirect effect0.0140.0050.0290.006 Living environment-social adaptationDirect effect0.0400.0150.0690.023– Social adaptation-physical healthDirect effect0.0430.021.0.0790.009– Social adaptation-mental healthDirect effect0.3490.2820.4040.000– Table 4Comparison of the model paths in different gender groups.Model pathEffect typeEffect estimate95% CIP-valueMediation test resultLowerUpperMale group Living environment-social adaptation-physical healthTotal effect0.0680.0380.0870.000Partial mediating effectDirect effect0.0630.0300.0920.000Indirect effect0.0050.0010.0110.038 Living environment-social adaptation-mental healthTotal effect0.2000.1540.2360.000Partial mediating effectDirect effect0.1820.1430.2250.000Indirect effect0.0180.0090.0340.008 Living environment-social adaptationDirect effect0.0560.0230.0790.005– Social adaptation-physical healthDirect effect0.0880.0480.1270.000– Social adaptation-mental healthDirect effect0.3300.2660.3830.000–Female group Living environment-social adaptation-physical healthTotal effect0.0420.0150.0720.007No mediating effectDirect effect0.0410.0130.0710.009Indirect effect0.001− 0.0010.0030.281 Living environment-social adaptation-mental healthTotal effect0.1740.1270.2110.000Partial mediating effectDirect effect0.1600.1190.2010.000Indirect effect0.0140.0050.0290.006 Living environment-social adaptationDirect effect0.0400.0150.0690.023– Social adaptation-physical healthDirect effect0.0430.021.0.0790.009– Social adaptation-mental healthDirect effect0.3490.2820.4040.000– Comparison of the model paths in different gender groups. In summary, comparative analysis across gender groups reveals an important differential pattern: the positive associations between the living environment and both physical and mental health are significantly stronger among older male adults in rural areas compared to females. This finding suggests that the health status of older men shows a stronger correlation with changes in environmental factors than that of older women. Further mechanistic tests confirm that social adaptation serves as a significant mediator in both groups, and notably, this mediating effect is more pronounced among older males. Collectively, these results highlight the presence of gender-based heterogeneity in the health benefits associated with improvements in rural living environments. Therefore, to maximize the effectiveness of health promotion efforts, future policies and practices should adopt a more targeted approach. Discussion Based on the social ecology theory, this study constructed an analytical framework of “living environment—social adaptation—physical and mental health” using a sample of 9,998 older adults from the CLASS 2023 survey conducted by Renmin University of China. Through SEM, it revealed the complex mechanisms of environmental impacts on older adults’ health and their gender-differentiated characteristics. This research not only provides a theoretical foundation for age-friendly environment renovation in practice but also expands the applicability of the Social Ecology Model in the field of gerontology. In terms of living environment, the study found that Chinese communities generally exhibit relatively high environmental quality, particularly in road traffic. However, certain shortcomings exist, such as obvious deficiencies in age-appropriate facilities, which require attention. In terms of health, the study revealed that older adults face challenges in both physical and mental health, with mental health issues being more severe. In social adaptation, older adults also demonstrated poor performance, with a lower acceptance level toward fresh things. More importantly, female older adults were found to be more vulnerable and disadvantaged in both physical and mental health and social adaptation, which is consistent with previous research conclusions39. Therefore, targeted improvements in the physical and mental health among older adults of different genders through living environment deserve in-depth exploration. First, the analysis of dual pathways linking living environment to older adults’ health confirmed a significant positive association between the living environment and both physical and mental health, with a notably stronger relationship observed with mental health. This is consistent with Gifford’s “environmental psychological perception priority” hypothesis40. Social adaptation was found to partially mediate the environment–health relationship, suggesting that the living environment is not only directly correlated with health through material improvements but also indirectly linked via enhanced social adaptability. This may be attributable to the fact that community environments rich in social participation opportunities, positive neighborly relations, and age-friendly facilities are associated with stronger social connectedness and adaptive capacity among older adults41. Improved social adaptation, in turn, correlates with reduced loneliness and anxiety, and helps maintain physical function through active social engagement, ultimately contributing to better mental and physical health29. These findings broaden the applicative scope of the Socioecological Model and affirm the mediating function of individual-level behaviors between macro-level environmental systems and health outcomes. Moreover, this study reveals significant gender-specific patterns through multi-group analysis, indicating that the positive associations between the living environment and both physical and mental health, as well as social adaptation, are more pronounced among older male adults. Grounded in socio-ecological theory and contextualized within the Chinese cultural setting, these differences stem from a differentiated interactive logic between multi-dimensional environmental factors and gender roles. Specifically: First, at the macro socio-cultural level, the traditional gender role of “men being responsible for outside affairs and women for domestic ones” continues to exert a profound influence42. Male older adults are often assigned the role of “primary breadwinner,” with their social value more closely tied to external, functional support from the environment. These elements are directly linked to men’s ability to stably engage in social production, and their social adaptation is closely correlated with self-identity and psychological well-being. Second, at the meso community and organizational level, current resource allocation and support mechanisms in China still exhibit certain gender biases43. Men often have greater access to environmental empowerment through community public services and are better positioned to convert such resources into social adaptation capacity. In contrast, social adaptation among women relies more heavily on emotional support from the community. However, such support tends to be less effective in enhancing core dimensions of social adaptation. Moreover, implicit gender discrimination in society further weakens the positive association between the living environment and social adaptation among older women. Third, at the micro family level, intra-household division of labor exacerbates these disparities. Men typically bear less responsibility for reproductive labor such as housework and intergenerational care. Improvements in the living environment can directly alleviate their survival pressure, allowing them to focus more on social participation. Women, however, often need to balance dual roles both inside and outside the home. Even with an improved living environment, their time and energy for social participation are frequently constrained by caregiving responsibilities, diluting the empowering effect of environmental improvements on social adaptation and thereby attenuating its ultimate correlation with health outcomes. This study enriches the gender-comparative perspective in aging and health research. It not only validates the differential applicability of socio-ecological theory across gender groups, but also provides policy insights for precisely enhancing health outcomes among older adults of different genders through living environment interventions within the Chinese context. This study has three limitations: first, although this study—based on cross-sectional data—can effectively examine the associations and mediating effects specified in the theoretical model, it does not establish clear causal relationships. Future research should therefore employ panel data to further validate these causal relationships. Moreover, the CLASS 2023 data capture a specific social context in the post-pandemic era, and the findings reflect the relationships among variables during this particular period. Consequently, future studies should incorporate longitudinal tracking to better understand the dynamic evolution of these relationships over time. Second, the assessment of the living environment in this study relied on subjective evaluations. Subsequent studies would benefit from integrating subjective appraisal systems with objective indicators in a scientifically systematic manner to enable a more comprehensive and in-depth exploration of the relationship between the living environment and the health of older adults. Third, while the sample covers 28 provinces nationwide, there are still areas not involved in the survey, In addition, analyses of group differences—such as urban versus rural contexts and socioeconomic disparities—were not sufficiently detailed. Future studies could further stratify older adult populations to better understand the differential mechanisms underlying environmental health effects across demographic subgroups. Conclusion The study found that older adults face challenges in both physical and mental health, with mental health being more vulnerable, and the health of female older adults is significantly weaker than that of males. This study, based on a “living environment—social adaptation—physical and mental health” framework, demonstrates that both the living environment and social adaptation are significantly and positively associated with the physical and mental health of older adults. Moreover, social adaptation serves as a partial mediator in the pathway linking the living environment to health outcomes. Notably, gender differences were observed in these associations: the living environment showed stronger correlations with physical and mental health, as well as with social adaptation, among male older adults compared to females. Therefore, it is necessary to provide targeted age-friendly environment renovation suggestions and strategies based on the needs of different older adult groups. The conclusions of this manuscript not only expand the theoretical exploration of the social ecology model in the field of health planning but also provide a basis for current age-friendly environment renovation practices in China. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1