the inuence mechanism of community-built environment on the health of the elderly: from the perspective of low-income groups

Background(cid:0)The international community has been paying attention to the health problems of the elderly and the age-friendly community of the elderly as the population ages, but there has not been enough discussion about the internal mechanism of the community-built environment to inuence the health of the elderly. Methods: In this study, descriptive statistical analysis, and structural equation Modeling (SEM) were used to make a group comparison between the elderly of different income groups. The data from this study, came from a sample survey in Shanghai, China. The study investigated the complex relationship among the community-built environment, social participation, outdoor exercise, and the health of the elderly, with a focus on the differences between in the elderly with different incomes level. Results(cid:0) The study found that health difference exists among the elderly in China: the lower the income, the worse the living environment, the worse the health. Community built environment has an important impact on the health of the elderly. And the community-built environment inuences the elderly's health through the intermediary role of outdoor exercise and social participation. Furthermore, the lower the elderly's income level, the stronger the direct effect of the community-built environment on their health; the higher the elderly's income level, the stronger the mediating effect of outdoor exercise and social participation on the impact of the community-built environment on their health. Community built environment play a more important role on the low-income elderly healthy. Conclusion: Governments should pay more attention to the health and living conditions of low-income elderly and take proactive steps to help them. Community design and building should pay more attention to the demands of low-income elderly groups, which will help to improve the health inequality of the elderly, consequently enhancing their overall health.


Background
Population aging is the current global population problem. The health problems of the elderly not only effect the solution of pensions and economic expenditures, but also determines the healthy level and quality of life in each country and even the global population. Since the publication of The Black Report (Black 1980Black et al. 1999), health inequalities has been an important topic in the international academic community (Marmot, 2005Scambler 2012Zhou et al., 2016. Health disparities exist objectively. The health status of higher socioeconomic groups is often better than that of lower socioeconomic groups, which is a phenomenon of the health inequality (Braveman 2006Claussen 2015Nurujeter et al 2018. The health equality of the elderly has been extensively concerned as the aging process accelerates. Therefore, it is important to pay more attention to the health issues that vulnerable elderly groups face, especially low income. From "successful aging" to "healthy aging" to "active aging", the concept of international community to deal with population aging has changed. The role of the elderly should be changed from "passive recipients of support" to "active participants in social activities" (WHO 2002). Then the concept of "age-friendly communities" emerged (WHO, 2007), with the ultimate objective of promoting more social activities and increasing health possibilities for the elderly (Menec Nowicki 2014Scharlach 2017. Because the community optimization construction has the characteristics of intervene ability and implementation, it has a lot of practical value for age-friendly communities to promote active aging (Plouffe & Kalache, 2010). The community is the basic activity places and living spaces. As a result, constructing a community-built environment that can effectively improve the health of various social groups has become a vital to promote social equity and improve the health of whole people. With the development of the practice of the age-friendly communities, international community, scholars, and governments are paying more and more attention to Nevertheless, the complicated interaction among the community-built environment, social participation, outdoor exercise, and the elderly's health has not been adequately explored.
Furthermore, in the context of a rapidly aging population, ignoring the difference between elderly groups (zheng et al., 2021), designing and constructing community-built environment through homogeneous and extensive measures will be the opposite of "age-friendly community" (zheng et al., 2020; zheng et al., 2021). To improve the overall health level of the elderly, we must recognize the diverse demands of different elderly groups, especially low-income groups. The elderly with various income levels showed signi cant differences in the living conditions, behavior habits and psychological needs, as well as the health factors. Therefore, the comparative study about the impact of the community-built environment on the health of elderly not only contributes to the overall promotion of the elderly's health, but also aids in the implementation of the concept of shared development in the eld of environmental research, as well as promotes social equity and social harmony.
Our research focuses on the living environment and health of the elderly in China, particularly in low-income groups. This is due to China recently facing the largest and fastest-growing aging population of the world (WHO,2012) and the large low-income group of the elderly. Our research primarily raises the following questions based on the above literature review and analysis: To make the sample as representative as possible, community samples should be selected as far as possible to cover different quality communities in Xinhua street. Due to geographical location, convenient transportation, and construction age, which often have a strong explanatory signi cation to the quality of the community. Therefore, this survey's primary sampling principle was based on the diversity of these aspects. See Figure 2 for details. Then, in the selected community, a sample survey of the elderly aged 60 and over without cognitive impairment was conducted.
The sample principle was if the number of elderly people in selected community was less than 120, all of them were surveyed; if the number of elderly people in selected community was more than 120, 120 elderly people were chosen at random. The list of the elderly without cognitive impairment was provided by the neighborhood committee, and 2783 valid samples were obtained. There were 1292 low-income elderly samples, 964 middle-income elderly samples and 527 higher-income elderly samples.

Measurement
Dependent Variable: The health of the elderly Self-evaluation of health has been widely used in self-perceived overall health (Jylhä, 2009Pagotto et al., 2013, which have highly predictive of functional disability, morbidity and mortality (Tsai et al., 2014) and even more important than actual medical measurements results (Maddox & Douglass, 1973). So, selfevaluation of health was thought an excellent predictor of objective health (Wu et al., 2013). This paper used self-rated health and health satisfaction to assess the health of the elderly. On a scale of 1 to 10, the higher the score, the better the health.

Independent variable: Community built Environment
Community-built environment includes leisure environment and landscape environment. They were based on two measurement models of the community-aware environment development by Mujahid et Al (2007). Leisure environment includes seven dimensions: walking convenient, walking tness, su cient trees, exercise opportunities, sports facilities, walking attraction and exercise attraction. And landscape environment includes three dimensions: the interest of architecture, environment cleanliness and the attraction degree. The responses to each item ranged from 1 to 5 (1=completely, 2=disagree, 3=neutral, 4=completely, 5=agree) and the higher score indicated higher degree of acceptance of the walking support environment and sensory support environment.
Intermediary Variables: Social Participation, Outdoor exercise The social participation of the elderly in this paper mainly refers to the activities of the elderly in the community, included ve activities: volunteer work, self-management and mutual assistance activities, lectures and reports, participation in cultural activities, and participation in interest groups. The elderly's level of social participation was assessed by asking them about the frequency of participated in various activities over the past 12 months. The item was scored on a scale 1 to 5(1=never, 2=several times a year, 3= several times a month, 4= once a week, 5=2-3 times a week), with the higher score indicating more social participation.
Outdoor exercise includes two observation variables: walking frequency and walking duration. The walking frequency was measured by the number of walking per week. Walking duration was about the time of each walk.

Control Variable
Age, gender, education and community residence time were included as control variable in this paper's conceptual model. Gender as 2 categorical variables, male as 0, female as 1. Education levels are assigned as follows: 1=junior high school and below, 2=senior high school, technical secondary school and technical school, 3=junior college, 4=bachelor, 5=master and above.

Statistical Analysis
This study analyzed how the community-built environment effect the health of the elderly through outdoor exercise and social participation as the intermediary. Structural Equation Modeling (SEM) is more suitable for the analysis of this complex relationship. The SEM has bene ts of visualization, intuition, and science in dealing with the comparative analysis of multi-group models. So, SEM and Maximum likelihood estimation method were used in this research. The structural equation model was tted using MPLUS software.
Multi-factor con rmatory analysis was performed on all the measurement models in the conceptual model, and the compositional reliability of all the measurement models was greater than 0.6; the average variance extraction was greater than 0.5; the factor load of the observed variables was greater than 0.6; the reliability coe cient was greater than 0.36 (Fornell & Larcker, 1981) and all the measurement models had good reliability and validity. The community sensory support environment and community walking support environment had a correlation coe cient of 0.632. Therefore, the community sensory support environment and the community walking support environment constituted the second-order model of the community-built environment.
The results of model tting demonstrate that the CFI did not achieve the ideal standard, indicating that the model had to be improved. after establishing the converged relationship between "sport facilities" and "exercise facilities", "interesting design" and "attractive", and "attractive" and "clean and tidy", the nal IFI, CFI and X2/DF all achieved the criteria thereby the optimized model was t. The nal indexes (CFI>.90, TLI>.90, RMSER<.08) achieved the criteria, which show that the model was t.

Results
Descriptive statistics The descriptive statistics of variables in Table 1 revealed that the elderly's health satisfaction was greater than self-assessment health. It indicated that the whole elderly had a better mentality and the self-assessment health and health satisfaction improved with income increased. In the community-built environment, the average of community leisure environment was often greater than the average of the community landscape environment. The mean value of all measurement variables in community-built environment re ected that the low-income elderly have lower than middle-income and high-income elderly. The low level of the elderly's social participation and the average of all participation activities were both below 2, implying that the elderly's participation occurred several times a year. Walking frequency was 4.2 times a week with a walking time of 28.57 minutes; walking frequency and time increased as income increased. In the control variables, the average age of the elderly was 72.7 years, with balanced the gender structure, and the overall education level was above senior high school, having a more than 22 years' living time in average. With rising income, the average age of the old reduced, the level of education gradually increased, the number of men climbed, and the living time reduced. The study applied the SEM latent mean comparison approach to compare the community-built environment, social participation, outdoor exercise, and health status of the elderly with different income levels. The processing of the mean of latent variables was one advantage of the analysis method. Unlike other statistical methods that add the mean to latent variables, the structural equation model was systematically analyzed by the different weights of each measured variable and eventually appears as difference among the means of variables in different groups. MPLUS set the low-income group to 0 and used software analysis to determine the speci c difference between middle-income group, high-income group, and low-income group. The use of a SEM latent mean comparison allows for a more precise measurement of the differences in variable means between various income groups ( Figure 2). The results of the model tting based on the entire sample are shown in Table 2 Table 3 Comparison of different income elderly model paths

Discussion
Our study explored the complex interaction among the community-built environment, social participation, outdoor exercise, and elderly health, as well as the difference between the elderly in different income groups, with a focus on low-income groups.
Our study con rmed the existence of health inequality problem (Braveman 2006Claussen 2015Nurujeter et al 2018 in the elderly. The greater income, the better health, and the living environment. The lower the income, the worse the health and living environment. The research also discovered that there are signi cant differences in the behavior of the elderly with difference income level: the higher the elderly's income, the lower the frequency of social participation. However, outdoor exercise showed an inverted V-shaped relationship, with the highest outdoor sports intensity in middle-income elderly and the lowest outdoor exercise intensity in the low-income elderly.
Our study con rmed the community-built environment had a signi cant impact on the health of the elderly (Menec & Nowicki, 2014;Moore, 2014;. Meanwhile, the community-built environment in uenced the health of the elderly through the mediation of the outdoor exercise and social participation.
That is, improving the quality of the community environment would increase the elderly's frequency of outdoor exercise and social participation, then improving their health.
More importantly, our study found signi cant differences in the pathways by which community-built environments affect the health of the elderly at different income levels. The lower the elderly's income level, the greater the direct impact of the community-built environment on their health. The higher the elderly's income level, the stronger the intermediary impact of outdoor exercise and social participation on the effect of the community-built environment on their heath. The community-built environment had a strong and direct impact on the health of the low-income elderly, and it was not in uenced by their behavior. While the impact of the community-built environment on the health of the middle-income elderly and the high-income elderly should be achieved through the intermediary impact of outdoor exercise and social participation.
The in uence path of community-built environment on health of the elderly with different income levels is different. Therefore, to truly improve the health of the elderly and reduce the health inequality, it was necessary to consider the needs of various income groups, with a particular focus on the characteristics of the low-income elderly, then achieve differentiated responses and accurate environmental governance. The community-built environment had an extremely important in uence in the health of low-income elderly, with a total effect value of 0.362. At the same time, this in uence was independent existence, will not be affected by behavior. The low-income elderly group had worst health status, and the community-built environment had the largest impact on the health of the low-income elderly, implying that improving their community-built environment would have the biggest impact on their health. Therefore, during the urban redevelopment process, relevant government departments and environmental designers should give special attention to the improving of the community-built environment of the low-income elderly.
However, the study still has some limitation. First, the survey scope and neighborhood sample quantity are limited. Since only Xinhua Community in Changning District of Shanghai was selected for the in-depth survey, research conclusion cannot represent all neighborhood environment of urban China and more empirical studies should be conducted in the future. Secondly, the representativeness of elderly samples requires further improvement. In terms of the selection of neighborhoods, although the current study takes geographical location diversity, transportation convenience and completion year as the sampling principles, it still fails to develop systematic random sampling. So, it needs to increase the uncertainty of elderly samples. Finally, since neighborhood environment in this study based on subjective assessments, follow-up research should combine subjective and objective system of neighborhood environmental assessments so as to better explore the association between neighborhood environment and health of the elderly.

Conclusion
The result show that there were issues of health inequality among the elderly. The higher the income, the better the health, and the lower the income, the worse the health status. As a result, extra attention must be paid to the health problems of the low-income elderly.
Our study shows that the community-built environment had a signi cant impact on the health of the elderly, and that behavior was an intermediary variable for the community-built environment to affect the health of the elderly. More importantly, we discovered differences in the impact of community-built environment on the health of different income groups. The lower the income level of the elderly, the stronger the direct effect of community-built environment on their health. The higher the income level of the elderly, the stronger the mediating effect of outdoor sports and social participation on their health.
We advise all governments to pay more attention to the health and living environment of low-income elderly people. We strongly suggest that, in the future planning, design, construction and renewal process of the community-built environment, more attention be paid to the needs of low-income elderly groups and that care be shown for them, which would assist to reduce health inequality. As a result, the elderly's health will improve.

Consent for publication
Not applicable Availability of data and materials The data and samples in this study were collected by Fudan University. The ethical approval code number is IRB#2015-12-0574. Due to the problems related to the use right, if necessary, please contact the author of the communication. Comparison of the community-built environment, social participation, outdoor exercise, and health differences between the elderly with different incomes Analysis Based on the Models of Full Sample Figure 3 The standardization coe cient of the path of the overall model for the elderly