Community built environment attributes moderate the relationship between family support and depression among older adults in urban China

Abstract

Background

This study examined the moderating effects of built environment attributes on the relationship between family support and depressive symptoms among community-dwelling older adults in urban China.

Methods

Quota sampling was used to recruit participants from Tianjin (one of the four municipalities in China) and Shijiazhuang (the capital city of Hebei Province). Face-to-face interviews were conducted with 799 respondents aged 60 years and older, in either their homes or local community centers. Multi-level modeling was used to test the proposed model.

Results

This study found that family support was negatively associated with depressive symptoms. Green spaces were negatively associated with depressive symptoms among older participants, although most indicators of the objective built environment attributes were not. The moderation analysis revealed a significant interaction effect between family support and green spaces on depressive symptoms. Specifically, living in areas with a higher percentage of green spaces may mitigate the negative effects of lower levels of family support on depression.

Conclusions

This study contributes a new direction for investigating the relationship between family support and depressive symptoms among community-dwelling older adults in China by considering the moderating effect of objective built environment attributes. The findings may guide practices and urban design in mental health promotion for older adults. Specifically, this study provides evidence useful for both policy designers and urban planners by highlighting modifiable environmental and objective factors that can promote community mental health for older adults who find it difficult to obtain family support in modern society.

Background

According to the latest data from the World Health Organization, the global prevalence of depression among older adults was 3.8% in 2021 [1]. Together with the intense trend of an aging population, depression has become a public health issue in China, with approximately 36.9% of older adults reporting depressive symptoms [2]. Many studies have documented that depression is a chronic and easily recurrent mental disorder that can considerably impair physical function and cause a high risk of suicide and mortality among older adults [3,4,5]. Moreover, it notably influences older adults’ social engagement with their families and communities [6].

According to a systematic review, embedding oneself in a network or receiving social support from friends reduces the risk of depression [7]. Compared to older adults in Western countries, older adults in China are more affected by family support [1]. Considering the dominant role of filial culture in Chinese communities, family support particularly informs older people’s sense that they are being cared for, that others are concerned about their well-being, and that they are respected [8], which prevents the occurrence of depressive symptoms.

However, given the trends of urbanization and family nucleation, community-dwelling older adults may experience difficulties in receiving family support. Hence, local communities have become a core area for older adults who typically spend more time in, particularly those with physical dysfunctions and limited mobility. More evidence on the community environmental correlates of mental health is needed. However, studies exploring family support and depression among older adults seldom consider community objective built environment attributes [9,10,11]. To increase the effectiveness of mental health–related evidence for policymakers and social service planners, it is essential to identify modifiable factors related to objective built environment attributes and their interaction effects with familial factors that can impact the risk of depressive symptoms among older adults. The present study attempted to address this gap in the accumulated knowledge by examining the moderating effect of objective built environment attributes on family support and depressive symptoms among community-dwelling older adults in urban China.

Family support and depressive symptoms among older adults

Although both quantitative and qualitative measures of family support can have positive effects on mental health [12], most scholars have found that perceived support is a better predictor of older adults’ psychological status [13,14,15]. Notably, perceived social support can be increased and depression can be alleviated in older individuals [10] through high-quality support from family members that nurtures feelings of concern, trust, and respect [8]. The main supportive role of families in aged care is highlighted and reinforced by both filial culture and the legal system in China. Regarding depression, which is generally underdiagnosed and undertreated (due to social stigma and misdiagnosis), family support is the most effective way to reduce depressive symptoms in older adults within China’s underdeveloped long-term care system [9, 11]. However, disassociation from the traditional family, the “empty nest” phenomenon, and the weakening of family ties are becoming more common in China, making it difficult to sustain family support. This may explain, to some extent, the recent higher prevalence of depression among older adults compared with that found a few decades ago in China [11]. Therefore, a more complete understanding of depression in later life within a non-familial context and its interaction with family support is required.

The moderating role of objective built environment attributes

Considering the widespread adoption of “aging in place” policy, the influence of objective built environment attributes on depressive symptoms may be particularly pertinent for older people. Some scholars have reported that neighborhood access to public transport, services, and amenities provides older adults with opportunities for an active social lifestyle that may reduce their risk of depression [16]. Similarly, research adopting recent national datasets in China indicated that older adults living in communities with better objective facilities (e.g., healthcare and consumption resources, exercise and recreational facilities, and religious and ancestral worship buildings) tended to have lower rates of depression [17].

Nevertheless, research and practice have not yet considered the integration of family support factors and objective built environment attributes. Previous studies have mostly focused on how family supportive roles [9, 11] and objective built environment attributes [16, 17] can prevent depressive symptoms in older adults and ignored the complex dynamics between the family and ecological environment.

We propose that objective built environment attributes can moderate the relationship between family support and depression among older adults in China. According to ecological systems theory [18], individuals’ mentalities, behaviors, and functions are shaped by interactions between personal and environmental systems. which helps explain the nature and outcome of the interactions between people and their support systems. This theory looks at individuals’ development across various levels of the community environment, namely, the microsystem, mesosystem, exosystem, macrosystem, and cronosystem. The microsystem refers to relationships and interactions individuals have within their immediate surroundings, such as within their families and neighborhoods. The mesosystem connects the structures within the individuals’ microsystem. The exosystem encompasses the social system, with which individuals do not directly interact. The macrosystem is the outermost layer of the individual’s environment and is characterized by cultural values, customs, and laws. The chronosystem system involves the relationship between the dimension of time and the individual’s environments [19]. The person-in-environment framework informs this study’s understanding of the association between family support and mental health among older adults. Strong filial traditions in Chinese macrosystems [9, 11] may cause Chinese older adults’ support systems (microsystems) to significantly shape their well-being. Furthermore, as indicated by previous research, the physical community environment is a protective factor against depression among older adults [20], especially for older individuals living alone, who may be more dependent on their community’s physical environment for services, amenities, and social interactions [21]. Building on previous research and theoretical frameworks, this study aimed to understand how family support is associated with depression among community-dwelling older adults in urban China by considering the potential moderating effect of objective built environment attributes. The proposed conceptual framework is presented in Fig. 1.

Fig. 1

Conceptual framework

Methods

Research design

Data for this study were derived from a survey “Social Capital, Intergenerational Solidarity, and Mental Health among Older Chinese Adults” which was jointly conducted by the Hebei University of Economics and Business and Renmin University from 2020 to 2023. Research participants were selected from Shijiazhuang, the capital of Hebei province of mainland China, and Tianjin (one of the municipalities under the direct administration of the State Council of China alongside Beijing, Shanghai, and Chongqing). These two cities adjacent to Beijing and are considered as one of the most important economic engines of mainland China. The sampling, measurements, and data analysis process were as follows.

Data and sampling

We determined the sample size for this study as follows: first, we checked a systematic review on social support and depression among community-dwelling older adults in Asia and found that the mean effect size of the reviewed studies was 0.35 [3]. Second, we calculated the sample size using the G-Power computer program with an effect size of 0.35, required power level of 0.95, and number of predictors of 19 (including independent and control variables). The results showed that at least 103 participants were required to detect moderate effects using linear multiple regression. Finally, as this study was supported by the Project of the National Social Science Foundation of China, we were expected to use a large sample size. Some researchers have recommended that a large sample should have at least 800 cases [22, 23]. Therefore, we determine the sample size for this study as 800.

Quota sampling was used to recruit participants from Tianjin and Shijiazhuang. First, five districts from each city were randomly selected. Second, two communities were randomly selected from the designated districts; in total, the study included 20 communities. In each community, 40 respondents aged 60 years or older were invited to participate in the study. The age and sex ratios of the participants were strictly controlled in accordance with the statistical representation in the latest local population census data. During the recruitment process, if the number of interviewees in one age or sex group was enough, we would skip the participants. In addition, older adults who were invited to participate in the study had to meet the following criteria: (1) have a local household registration status, (2) aged 60 years or older, and (3) have lived in their local community for more than 180 days in the previous 12 months.

Face-to-face interviews were conducted at local community centers or at the homes of the participants by trained interviewers. Informed consent was obtained from all the respondents before the interview. The participants were also informed of their right to withdraw from the interviews. The Short Portable Mental Status Questionnaire was used as a screening tool to assess cognitive function [24]. A total of 853 people were interviewed; 800 completed the interviews. Our study was approved by the Institutional Review Board (Reference No.: EA2003026). After removing cases with missing values for key variables, a total sample of 799 individuals were included in the analysis.

Measurements

Dependent variable

Depression was the dependent variable in this study. We used the 10-item Center for Epidemiologic Studies Depression Scale (CES-D-10), which is commonly used to measure depression in older respondents [25, 26]. This scale included negative-mood items, such as “I am troubled by some small things” and “I feel lonely”; items related to positive emotions, such as “I am happy” and “I am hopeful for the future”; and items related to somatic syndromes, such as “My sleep is not good.” Responses were assessed on a 5-point Likert-type scale (1 = rarely or none of the time, 2 = not much of the time, 3 = almost half of the time, 4 = most of the time, and 5 = almost every day). Two reverse questions were reverse coded. Summed scores represent the level of depressive symptoms (in the range of 0–50), with higher scores indicating higher levels of depression. The Cronbach’s alpha of the CES-D-10 was 0.755 in this sample, indicating good reliability of the measurement [27].

Independent variable

Family support was the independent variable. It was measured using the 4-item Multidimensional Scale of Perceived Social Support [12, 28]. Respondents were asked whether they agreed with the following four statements: (1) “My family really tries to help me”; (2) “I get emotional support from my family”; (3) “I can discuss my problems with my family”; and (4) “My family is willing to help me make decisions on important issues.” Respondents indicated their answers on a 5-point Likert scale (1 = strongly disagree, 3 = neutral, and 5 = strongly agree). Average scores were calculated, with higher scores indicating higher quality family relationships. The Cronbach’s α for this scale was 0.846.

Moderators

Objective built environment attributes were the moderators in this study. The objective built environment attributes were measured at the 20 community level. Following the 3D framework suggested by Renalds [29], six built environment attributes in three “D” dimensions—“Design” (i.e., street connectivity, green space), “Diversity” (i.e., entropy land use mixed index), and “Destination” (i.e., availability of recreational, shopping, and community services)—were measured within a 1 km buffer area from the centroid of the community center. Street connectivity was calculated as the number of road intersections divided per square kilometer (i.e., number of road intersections (n)/area(s)). Green space, which refers to green features, including vegetation cover (e.g., grasslands, woodlands, and shrublands) but excluding agricultural land, was calculated as a percentage of green space area (i.e., green space divided by the buffer area) (i.e., % green space). The land-use mixed index was calculated using an entropy algorithm based on six land-use types (residential, commercial, industrial, institutional, open space, and recreational), ranging from 0 (homogeneous land use) to 1 (equal mix of land-use types) (i.e., ${\raise0.7ex\hbox{${\_\left( {\left[ {\sum\nolimits_{j = 1}^k {{P^j}\ln {P^j}} } \right]} \right)}$} \!\mathord{\left/{\vphantom {{\_\left( {\left[ {\sum\nolimits_{j = 1}^k {{P^j}\ln {P^j}} } \right]} \right)} {\ln k}}}\right.\kern-\nulldelimiterspace}\!\lower0.7ex\hbox{${\ln k}$}}$) [30]. Availability was measured as the number of services available in an area (i.e., number of services (n)). Recreational services included rest gardens, parks, playgrounds, sports grounds, indoor game halls, recreation centers, and sports centers. Shopping services included supermarkets. Community services included public government services. Python programming based on Jupyter platform was used to extract geographical information from OpenStreetMap, which includes the open-source mapping recourses. The official lists of the above types of services were used for double checking with the sources collected from OpenStreetMap to ensure including all services. ArcGIS version 10.2 was used to create the built environment variables.

Control variables

The analysis controlled for the participants’ sociodemographic characteristics and health statuses, including age, sex, marital status, educational attainment, living alone, income, and functional health. Patients’ age and sex were self-reported. Marital status, living status, and level of educational attainment were classified dichotomously (0 = unmarried, 1 = married, 0 = living with others, 1 = living alone, 0 = primary school or lower, and 1 = secondary school or higher). Income was used as a continuous variable. Functional health was measured by Instrumental Activities of Daily Living [31]. Difficult level (i.e., 0 = no difficult, 1 = some difficult, 2 = very difficult) of seven daily activities (i.e., prepare meals, ordinary housework, property management, medication treatment, using the phone, going out shopping, getting out on a ride) were summed, with the higher score referring to higher difficult (i.e., lower functional health). The Cronbach’s α for this scale was 0.785 in this study.

Statistical analysis

Hot-deck imputation was used to impute the missing data. Descriptive statistics were calculated for the sample characteristics and objective built environment attributes of the communities. Multilevel modeling was performed to test whether the built environment moderated the association between family support and depression.

The values of the variance inflation factor (VIF; VIF values above 10 indicated multicollinearity) of the independent variables in the present regression model ranged from 0.00 to 2.79, indicating the absence of multicollinearity [32]. The normal distribution of the key variables was checked using skewness and kurtosis. The skewness and kurtosis of depression were 1.214 and 1.045, respectively. The skewness and kurtosis of family support were − 1.296 and 3.195, respectively. The skewness of street connectivity, green space, the land use mixed index, number of recreational services, number of shopping services and number of community services were 1.923, 0.014, 1.674, 1.743, 1.645, and 0.987, respectively; the kurtosis of street connectivity, green space, the land use mixed index, number of recreational services, number of shopping services and number of community services were 2.796, 2.687, 4.632, -0.964, 4.721, and − 1.033, respectively. Therefore, the key variables of this research are normally distributed [33, 34]. Outliers were screened using a multivariate outlier analysis; the Mahalanobis distance results revealed that no cases had significant D². None of the variables were missing more than 2.5% of their overall number of values, and those that were, followed a random pattern. According to the criteria proposed by Tabachnick and Fidell [35], missing less than 5% of data is acceptable. Missing values in this dataset were handled in the listwise deletion through the linear regression process.

In the first step, respondents’ sociodemographic characteristics were entered into the model. Second, family support and built environment variables were entered into the model after adjusting for individual sociodemographic characteristics. In the third step, we entered the two-way interaction terms of all built environment attributes and family support into the model to examine whether the associations between family support and depression varied by community built environment. We also conducted a stratified analysis by running separate models for samples residing in lower- and higher-built environment attributes, which were distinguished by the media of the variables. Data analyses were performed using R version 4.1.3.

Results

Sample characteristics

The characteristics of the sample population and the streets on which the respondents resided are presented in Table 1. This study included 799 participants with a mean age of 70.392 years (SD = 7.041). The majority of participants were aged 65–79 years (67.46%), were female (61.20%), were married (77.97%), had attended secondary school or higher (75.09%), lived with others (87.36%), and had an annual household income of more than 5000 RMB (mean = 5297.926). The mean vale of functional health was 0.567 (SD = 1.705). The mean scores for family support and depression were 4.476 (SD = 0.574) and 5.812 (SD = 5.303), respectively. For the objective built environment attributes of the 20 communities included in the analysis, street connectivity ranged from 10.321 to 39.252 road intersections per square kilometer (mean = 16.342, SD = 9.751), green space ranged from 0.9 to 29.5% (mean = 16.3%, SD = 0.084), and the land-use mixed index ranged from 0.427 to 0.659 (mean = 0.512, SD = 0.056). In the communities in which the participants lived, the mean of number of recreational services, shopping services, and community services were 4.256 (SD = 2.845), 5.326 (SD = 2.185), and 1.756 (SD = 1.689), respectively.

Table 1 Characteristics of the sample population (n = 799) and the streets (n = 20) on which participants resided

Interaction effects of family support and objective built environment attributes in depression

Table 2 presents a multi-level regression summary of the demographic factors, family support, objective built environment attributes, and interaction terms regressed on depression. Model 1 showed that age (β = 0.144, p <.05) and functional health (β = 0.692, p <.01) were significantly associated with depression, whereas Model 2 showed a significant relationship between age (β = 0.223, p <.05), functional health (β = 0.707, p <.01), family support (β = -0.095, p <.001), and green space (β = -0.514, p <.05). In Model 3, the interaction terms of family support and green space were significantly associated with depression (β = -0.923, p <.05). Thus, for objective built environment attributes, green space was a significant moderator in the relationship between family support and depression among community-dwelling older adults in China.

Table 2 Interaction effects of family support and objective built environment attributes on depression

Stratified analysis for family support and depression by objective built environment attributes

Table 3 shows the results of the stratified sensitivity analysis that compared the relationship between family support and depression between lower and higher environmental attributes. These results are consistent with the findings of the regression analysis with the interaction terms. This shows that family support and depression have a negative association pattern in most built environment areas, similar to that found in the general model. However, this negative association disappeared only in areas with high greenspace percentages. The interaction effect was visualized in Fig. 2. As it demonstrated, the lower line presented older adults living in areas with higher greenspace percentages, the effect of family support on depression became insignificant (β = -0.369, p =.965). For the other line, it shown negative association between family support and depression (β = -0.742, p <.001) among the group living with lower green space percentages.

Table 3 Stratified analysis for family support and depression by objective built environment attributes

Fig. 2

The relationship between family support and depression by green space

Discussion

Within the prevalence of policies that prioritize aging-in-place, our study is among the first to explore the association between family support and depression in community-dwelling older adults in urban China by considering the potential moderating effect of objective built environment attributes. This study has several key findings. Notably, we found that family support is negatively associated with depressive symptoms. Green spaces were negatively associated with depressive symptoms among older participants, although most indicators of the objective built environment attributes were not. Interestingly, based on the results of the moderation analysis, there was a significant interaction between family support and green space on depressive symptoms. Specifically, the results suggested that living in areas with a higher percentage of green space may mitigate the negative effects of lower levels of family support on depression.

In our study, we echoed the available literature [8, 9, 11, 13,14,15] by finding that family support acts as an important protector against depressive symptoms in Chinese communities influenced by the culture of family parochialism and filial piety, which frames older individuals’ meaning of life and mental health. However, as indicated in the Introduction, the protective function of family support has eroded with the rapid process of modernization in China. Therefore, social service delivery agencies and policy designers should identify modifiable factors within non-familial contexts.

In line with the theoretical grounds for emphasizing the importance of environmental attributes on the physical and mental health of residents, this study suggests that green spaces or other natural scenes are critical elements in restorative effects and can result in meaningful reductions in psychological distress. As suggested by the stress reduction theory (SRT), attention toward natural, unthreatening environments is said to demonstrate that humans are both “evolutionarily adaptive”—because natural elements were critical for early humans’ survival and well-being—and “biologically adaptive” [36, 37]. In addition, as indicated by the attention restoration theory (ART), natural environments may facilitate recovery from the mental fatigue caused by having to consistently focus during daily tasks [38, 39]. Exposure to green spaces is a quick-onset positive affective reaction that may contribute to the restoration of depleted cognitive capacity, enhance recovery from periods of psychosocial stress, and increase optimism [40]. Thus, we suggest that encouraging older adults’ exposure to natural scenes may be an effective option for improving their mental health through social service strategies.

This study contributes a new direction for investigating the relationship between family support and depressive symptoms among community-dwelling older adults in China by considering the moderating effect of objective built environment attributes. By identifying these characteristics, our results offer insights that can guide practices and urban design to promote mental health among older adults. Ultimately, this study suggests that living in areas with a higher percentage of green space can mitigate the negative effects of lower levels of family support on depression among older adults. It provides evidence for both policy designers and urban planners by highlighting modifiable environmental and objective factors for older adults who find it difficult to obtain family support in modern society to promote community mental health. This is in line with the World Health Organization’s Mental Health Action Plan 2013–2030, which advocated for a multisectoral approach toward the prevention of and enhanced recovery from mental illness, promotion of mental well-being, and reduction in disability and mortality among people living with mental disorders.

However, according to the 3D framework suggested by Renalds [29], we did not find evidence that other objective built environment attributes—namely: diversity, destination, and connectivity—were related to depression, in contrast to other studies [16, 17]. The first reason may be that the influence of built environment attributes on depression mostly involves mediating mechanisms. As indicated by a systematic review, associations between the physical features of community context and depression were less consistent than those related to social processes; this may be because built environment features affect social connections and social support among older residents, which in turn may affect their vulnerability to stress and depressive symptoms [41, 42]. Therefore, the insignificance may reflect that the present study only considers the physical properties of built environment attributes, and does not explore older adults’ use of and satisfaction with the built environment. Secondly, for many older adults in China, neighbors play important roles in shaping their mental health by giving them a sense of belonging and connectedness [43], which may lead them to be less psychologically sensitive to relational/objective opportunities present in the attributes of the built environment. Our results thus suggest that older adults may tend to under-use built environment attributes, such as diversity, destination, and connectivity, if they have already attained available resources from their nearest neighbors to preserve their mental health and maintain their social roles. Lastly, the differences between our findings and those of other studies conducted in China that indicate that older adults living in communities with more advantaged objective facilities tend to have lower rates of depression, such as a study by Lei and Feng [17], may be due to differences in the study sample. This study’s sample was limited to older adults living in urban communities, a population that possesses more homogeneous objective built environment attributes of diversity, destination, and connectivity characteristics. Lei and Feng [17] examined older adults in various living situations in both urban and rural areas.

Similarly, the moderating roles of diversity, destination, and connectivity in the relationship between family support and depression among older adults were not statistically significant. There are two possible explanations for why green spaces could modify the relationship between family support and depression among older adults, but other objective built environment attributes (diversity, destination, and connectivity) could not. First, the objective built environment attributes of diversity, destination, and connectivity could meet the daily needs of older adults, most likely by providing instrumental support for healthcare services, amenities, and social interactions. Based on ART [39] and SRT [37], nature and green spaces are not merely settings but can also be partners/supporters for reducing stress. They play an accompanying role for older individuals and provide them with support. Therefore, we can infer that older adults who lack family support in urban communities are more vulnerable to depressive symptoms because they are less accompanied and need more environmental attributes that offer accompanying support. In contrast to the other three attributes, green space was considered a significant and proximal protective factor against depression among older adults living in urban communities lacking family support. Moreover, diversity, destination, and connectivity may have indirect effects on depression through perception, cognitive evaluations, and behaviors, such as daily necessary purchases or physical activity, which may involve further complicated procedures. However, green spaces also directly affect depression, which is a simpler approach. This might be why it can be an easier moderator because, as long as people immerse in it rather than doing anything, it would start to play a role.

This study also had the following limitations. First, we did not use a random sampling method to recruit respondents. The data were collected via quota sampling. To gather a sample that appeared to be representative of the population, the age and sex structure of the participants was strictly maintained. The sampling selection process may have introduced bias, which may limit the empirical generalization of the findings. Second, we adopted an administrative boundary rather than creating a buffer area during the analysis process, which may involve a modifiable areal unit problem. Third, this study failed to compare the results from different geographical units because of the use of administrative boundaries and a lack of data on the exact address of residential locations, which may lead to uncertain geographic context problems. In future studies, the exact address can be collected and multiple buffers can be drawn to conduct a sensitivity analysis.

Conclusions

This study explored the moderating effects of objective built environment attributes on the relationship between family support and depressive symptoms among older adults living in urban communities in China. The results suggested that the older adults may mitigate the depressive effects of poor family support by living in areas with more green space.

This study is among the first to explore the relationship between family support factors and the objective built environment in the context of depression among older adults in China. With the advocacy of the basic concept of healthy ageing [44], an objective built environment has influentially shaped the dynamic process of interpersonal, social, and environmental interaction for older adults. This study offers insights on designing age-friendly communities that can promote mental health and maintain healthy ageing, which may be of interest to both researchers and practitioners.