Mitigating Health Disparities among the Elderly in China: An Analysis of the Roles of Social Security and Family Support from a Perspective Based on Relative Deprivation

Abstract

The joint involvement of family and society in elderly care is a crucial factor in improving the health status of older adults and narrowing health disparities, which are essential for achieving sustainable development goals. However, the interactions between these entities and their mechanisms of influence require further investigation. By utilizing data from the China Longitudinal Aging Social Survey (CLASS) spanning 2014 to 2016 and employing the Kakwani index of individual relative deprivation in conjunction with a two-way fixed-effects model for unbalanced panel data, in this study, we investigated the mechanisms through which social elderly care security and familial support influence health inequalities among the elderly. The findings reveal that only senior benefits (=−0.009, p < 0.05) significantly mitigate relative health deprivation in this population. Enrollment in pension insurance amplifies the sense of relative health deprivation among the elderly, but this effect becomes insignificant after controlling for temporal effects. Both economic support (=−0.002, p < 0.05) and emotional support (=−0.004, p < 0.01) from offspring significantly reduce the level of relative health deprivation among the elderly. Mechanism testing results indicate that individual attitudes towards aging serve as a mediator in the relationship between relative health deprivation and preferential treatment, economic support, and emotional support. The results of further heterogeneity tests suggest that the impact of various elderly support models on relative health deprivation differs by age, gender, and residential area.These findings confirm that support from both society and family plays a crucial role in achieving sustainable health outcomes for the elderly. Consequently, it is recommended to enhance the social elderly care security system, bolster familial support functions, cultivate positive individual attitudes towards aging, and address health inequalities among the elderly in accordance with their distinct characteristics, thereby improving their quality of life and sense of fulfillment, and contributing to the broader goals of sustainable development.

1. Introduction

The Chinese elderly population is facing unprecedented challenges in elderly care [1]. Against the backdrop of deepening economic structural reforms, large-scale urban migration, persistently low fertility rates, and increasing female labor force participation, family structures are shrinking and becoming increasingly nuclear. This reality, coupled with the growing geographical distance between elderly parents and their children, is continuously challenging traditional family care models [2]. In response to the trend of population aging, the Central Government and the State Council have implemented a series of plans and policies focusing on the “urgent, difficult, worrisome, and expected” livelihood issues of the elderly, aiming to construct a support and service system compatible with the aging process and continuously promoting the coordinated development of the aging cause and industry [3,4]. Despite the operation of the New Rural Pension Scheme launched in 2009, which integrated social elderly care into family care to compensate for its inadequacies, the level of protection it offers remains quite limited [5]. If the extension of elderly people’s lifespans is also accompanied by long-term illness, then the associated medical and care costs will pose a significant burden on families and society alike [6].

How can we ensure that more elderly people enjoy longevity and good health? In recent years, policymakers have prioritized creating a favorable social environment for elderly health. Policies such as the Healthy China 2030 Planning Outline and the 13th Five-Year Plan for the Development of National Aging-related Affairs and Construction of the Elderly Care System advocate improving health levels, enhancing health equity, and ensuring universal access to basic elderly care and public health services. Especially since the outbreak of the COVID-19 pandemic in 2019, there is growing consensus among all sectors on the need for integrating health factors into all policies, prioritizing people’s health in the development strategy, and refining relevant policies to comprehensively protect citizens’ health rights.

However, some scholars [7] argue that research on health should not be limited to macro-level differences in healthcare, resource allocation, and policies. It should also consider “inequality” among individuals at the micro-level. Despite significant improvements in the population’s health since the Chinese economic reform, known as the “reform and opening-up”, health inequalities remain pronounced across urban–rural divides, regions, and social strata [8]. When individuals exist within specific social structures and relationships, interactions between macro- and micro-factors lead to health disparities [9]. Especially in the context of transitioning elderly care models, the interrelations among family support, social security for the elderly, and health inequality become less clear-cut. In this study, we addressed the following questions: What is the relationship between family support and social security for the elderly regarding health inequalities? What mechanisms link traditional family support and continually improving social security with health disparities among the elderly? How do family support and social security impact health inequalities differently across genders, ages, and regions? Compared with existing studies, as presented here, we uniquely incorporated the individual relative deprivation index from a micro-perspective, assessed the synergistic and substitution effects of different elderly care models on health inequality among the elderly, and explored the micro-level mechanisms behind these effects. The findings offer valuable policy implications for enhancing the elderly care security system and improving the health standards of the elderly population.

2. Literature Review and Research Hypotheses

2.1. Social Security, Family Support, and Health Inequality (Relative Health Deprivation)

Compared with demographic changes, the transformation of family structures tends to be more complex. In the new era, on one hand, families seem to be in decline; on the other hand, within the context of a risk society, people’s reliance on family appears to have strengthened. Is there a potential for social elderly care to eventually supplant family care? Existing research has identified complex dynamics between family care and social elderly care systems, where the development of the latter can alleviate family care pressure and improve the quality of life for the elderly, potentially creating a “synergistic effect”. However, it may also diminish the role of family care to a certain degree, leading to a “substitution effect” [10,11]. Further research indicates that social care services can partially substitute family care, easing the burden on families. However, as the health of the elderly declines, the substitutive role of social services diminishes, and they cannot fully supplant family care [12]. The function of family-based elderly care should not be simply seen as diminished or deteriorating. When viewed from the perspectives of economic support and service provision, there is considerable variation in the economic support provided, while the service provision aspect is generally declining [13]. An increasing number of scholars are noting that the mode of elderly care is a critical factor influencing the health of the elderly. In the short term, social insurance has the potential to reduce health risks for the elderly under the positive regulation of elderly care services. The long-term positive impact of social insurance on elderly individuals’ psychological cognition is contingent upon the collaborative support of community elderly care services [14]. Community elderly care services not only encourage active physical engagement among the elderly, enhancing their independence and physical well-being, but also foster social interaction and cognitive stimulation, thereby improving psychological health [15]. Regarding family care support, both daily assistance and emotional comfort significantly benefit the health of the elderly, with daily life care having a more pronounced effect than financial support [16]. However, dissenting views from other studies suggest that the availability of pensions and the presence of community facilities for the elderly do not have a significant impact on self-rated health among the elderly. Aside from the economic support of children, other forms of family care support do not significantly impact self-rated health [17].

The elderly are in a phase of gradual physical decline, and within this demographic, there is clear health inequality favoring the wealthy [18]. With the economic and medical advancements post-reform and -opening-up, the Chinese public health system has rapidly improved. This has led to a shift in health consciousness among the lower social strata from a passive attitude to actively pursuing and maintaining health. Meanwhile, the upper strata has achieved a better health status by virtue of healthier lifestyles, less relative poverty, and better social medical security, further widening health inequalities among the elderly [19]. For instance, while pensions have improved the health status of the elderly and narrowed the health gap among the older and higher-pension-income individuals, they have widened the one among the middle-aged and lower-pension-income persons [20]. Health inequity and its correlation with social security have garnered wide attention. However, discussions on how social and family elderly support promote health equality, especially from a relative deprivation perspective, are still in the early stages. Here, we define health inequality as an individual’s disadvantaged position relative to others within a group, where greater relative health disadvantage leads to more severe health deprivation and higher levels of health inequality [21]. The perspective of relative deprivation provides an effective micro-level breakthrough for studying health inequality issues.

Hypothesis 1.

Elderly individuals who enjoy social security benefits and receive family support for the elderly experience lower levels of relative health deprivation.

2.2. Social Security, Family Support, and Individual Attitudes towards Aging

Existing research has found that intergenerational support provided by children has a significant positive effect on the elderly’s attitudes towards their own aging process, with the impact of different forms of support decreasing in the following order: emotional support, care giving, and financial support [22]. In rural areas, the more support the elderly receive from family and friends, the more positive and optimistic their attitudes are [23]. At the societal level, the higher the quality and degree of actual social support received from others, the more positive the attitudes towards aging [24]. Specifically, social security participation can positively influence attitudes towards aging, and children’s financial support and community care can mitigate negative ones [25].

Hypothesis 2.

Elderly individuals who benefit from social security for the elderly and receive family support have a more positive attitude towards aging.

2.3. Individual Attitudes toward Aging and Health Inequality (Relative Health Deprivation)

Levy’s (2009) [26] Stereotype Theory suggests that attitudes towards aging impact the health of the elderly through physiological, behavioral, and psychological pathways. The findings of domestic studies also show that elderly people have more positive than negative evaluations of aging; attitudes towards aging are closely related to psychological health. Ageist attitudes may lead to increased stress, anxiety, and depression and reduce the psychological well-being of the elderly. The internalization of negative stereotypes about aging can further exacerbate these issues, causing older adults to withdraw from social interactions and experience a decline in physical activity, thereby accelerating the deterioration of health [27]. The “Wilkinson Hypothesis” posits that as the overall income disparity in society increases, social cohesion gradually diminishes, leading to an increase in feelings of relative deprivation, which in turn can lead to a general decline in public health standards [28,29,30]. Health inequality is influenced by both the micro-level socio-economic characteristics of individuals and macro-level social policies, with various intermediate mechanisms contributing to health inequality [31]. Although direct empirical research on the relationship among social security for the elderly, family support for the elderly, and relative health deprivation is lacking, evidence suggests potential connections. For instance, a study on social isolation among the elderly found that those experiencing social isolation had worse health outcomes [32] (Figure 1).

Hypothesis 3.

The more positive an individual’s attitude towards aging, the lower their sense of relative health deprivation.

Hypothesis 4.

An individual’s attitude towards aging mediates the relationship between elderly support and relative health deprivation, meaning that elderly support from family and society leads to a more positive attitude towards aging, thereby reducing the sense of relative health deprivation.

3. Methodology and Data

3.1. Data Source

In this study, we utilized data from the 2014 and 2016 iterations of the “China Longitudinal Aging Social Survey” (CLASS), based on stratified multi-stage probability sampling among individuals aged 60 and above, encompassing 28 provinces (including autonomous regions and municipalities) across China, with the exception of Hong Kong, Macau, Taiwan, and the autonomous regions of Xinjiang and Tibet. The survey covers a wide array of topics, such as economics, marriage, family, eldercare services, and intergenerational relationships, facilitating the assessment of the impact of various social security policies for the elderly and familial intergenerational support on the health of the elderly population. After excluding cases with missing key variables, the final sample size was 16,337, comprising 48.19% males and 51.81% females; among these individuals, 56.04% were from urban areas, while 43.96% were from rural regions, and their average age was 70.51 years.

3.2. Model Construction and Variable Setting

To validate the research hypotheses stated earlier, eliminate the interference of time-invariant unobserved factors, and reduce the bias caused by endogeneity to ensure more reliable econometric results, in this study, we applied a two-way fixed-effects model using unbalanced panel data to control for individual and time effects. We investigated the effects of social pension security and family support on health inequality among the elderly. The baseline model is expressed as:

$$Relativehealt{h}_{it}={\beta }_0+{\beta }_{1}agingsecurit{y}_{it}+{\beta }_{2}familysuppor{t}_{it}+contro{l}_{it}+{\delta }_i+{\mu }_i+{\epsilon }_{it}$$

(1)

In Equation (1), the dependent variable, $Relativehealt{h}_{it}$, represents the relative health deprivation status of the i-th elderly respondent in year t; aging $securit{y}_{it}$ is the social security variable for the elderly, indicating the status of the social security benefits enjoyed by the i-th elderly respondent in year t; family $suppor{t}_{it}$ represents the family support for elderly care for the i-th respondent in year t; control stands for control variables; ${\delta }_i$ represents the intercept term for individual heterogeneity that does not vary over time; ${\mu }_i$ represents fixed effects over time; and ${\epsilon }_{it}$ is the error term that varies across individuals and over time.

To further examine the mediating effect of individual attitudes towards aging between different elderly support mechanisms and relative health deprivation among the elderly, the following mediation model was constructed:

$$Relativehealt{h}_{it}=c\times agingsecurit{y}_{it}+e_1$$

(2)

$$attitud{e}_{it}=a\times agingsecurit{y}_{it}+e_2$$

(3)

$$Relativehealt{h}_{it}=c^{\prime }\times agingsecurit{y}_{it}+b\times attitud{e}_{it}+e_3$$

(4)

$$Relativehealt{h}_{it}=c_1\times familysuppor{t}_{it}+e_4$$

(5)

$$attitud{e}_{it}=a_1\times familysuppor{t}_{it}+e_5$$

(6)

$$Relativehealt{h}_{it}=c_1^{\prime }\times familysuppor{t}_{it}+b_1\times attitud{e}_{it}+e_6$$

(7)

In this context, $attitud{e}_{it}$ represents the individual aging attitude of the i-th elderly person in year t, with a, b, c, $c^{\prime }$, etc., being the parameters needing estimation. Equations (2) and (5) represent the direct effects, Equations (3) and (6) the indirect effects, and Equations (4) and (7) the total effects. In this study, we employed the Sobel mediation test to examine the mechanisms by which social security for the elderly and intergenerational family support affect relative health deprivation among the elderly. By resampling the original sample multiple times, we calculated the coefficients of the effects and the proportion of the mediated effects and determined the significance of these effects by checking if the confidence intervals excluded zero.

3.2.1. Dependent Variable

Relative health deprivation served as the dependent variable in this study. To examine the impact of different elderly support mechanisms on relative health deprivation among the elderly, in this study, we selected the Kakwani index of individual relative deprivation to represent the dependent variable. Objectively, there are indeed differences in the physical health conditions of the elderly, with younger elderly individuals being healthier than older ones. However, subjective evaluations of health may vary due to differences in environmental and resource endowments. Furthermore, previous studies that have measured health inequality from psychological and physiological dimensions may have not fully captured the essence of “relative deprivation”. Building on research on this topic by Song, Shi (2022) [33] and Chunhai, Tao (2022) [34], in this study, we employed a revised Kakwani index based on self-rated health variables as an indicator for measuring individual relative health deprivation, as detailed below:

$$RD(y,y_i)={\gamma }_{yi}({\mathsf{\mu }}_{yi}-y_i)/{\mathsf{\mu }}_Y$$

(8)

where ${\mathsf{\mu }}_Y$ represents the average self-rated health of all individuals in group Y, ${\mathsf{\mu }}_{yi}$ is the average value of self-rated health that exceeds the sample health in group Y, and ${\gamma }_{yi}$ represents the percentage of the number of samples exceeding health in Y out of the total number of samples in Y (definitions and value ranges are shown in Table 2).

3.2.2. Core Explanatory Variables

The core explanatory variables are measured in two dimensions: social security for the elderly [35] and family support for the elderly [36]. For specific value assignment details, please refer to

Table 1. Settings of core independent variables.

Dimension	Variable Name	Indicator	Variable Assignment
Social pension security	Pension insurance	What social security benefits do you receive? (Rural insurance, urban employee insurance, urban resident insurance, and retirement pensions for government or public institution employees)	1 = received; 0 = not received
	Elderly care services	Have you used any of the following types of social eldercare services? (Home visits, housework, shopping assistance, medical accompaniment, senior service hotline, legal aid, senior meal or meal delivery, day care or eldercare station, and psychological counseling)	1 = used; 0 = not used
	Senior benefits	Have you enjoyed any local senior benefits (such as free park visits and public transportation)?	1 = enjoyed; 0 = not enjoyed
Family eldercare support	Financial support	In the past 12 months, has your child given you any money, food, or gifts? What was the total value of these items?	From 1 (nothing given) to 9 (more than or equal to CNY 12,000)
	Living support	In the past 12 months, how often has your child helped you with household chores?	Almost never = 1; a few times a year = 2; at least once a month = 3; at least once a week = 4; almost daily = 5
	Emotional support	In the past 12 months, how often have you called your child?

Source: Created based on the CLASS. Living support and emotional comfort were reverse-coded based on the original survey responses.

.

3.2.3. Mediating and Control Variables

Individual attitudes towards aging serves as the mediating variable in this study. It was measured with eight items: “If given the opportunity, I would like to participate in some work organized by the village/community committee”, “I often think about doing something for society”, “I now enjoy learning”, “I feel that I am still a person useful to society”, “Society changes too fast, and I find it hard to adapt to these changes”, “Nowadays, more and more viewpoints are difficult for me to accept”, “More and more of today’s new social policies are hard for me to accept”, and “The changes in today’s society are increasingly unfavorable to the elderly”. The answers were scaled as completely disagree = 1, somewhat disagree = 2, neutral = 3, somewhat agree = 4, and completely agree = 5. To standardize the scoring, the last four items were reverse-scored and then summed together, with a score range of 8–40 points. Higher scores indicate a more positive attitude towards aging (variable definitions and values are shown in Table 2).

3.2.4. Descriptive Statistics of Data

Table 2. The descriptive statistics of the variables.

Variable Name	Variable Definition and Values	Mean (2014)	Std. Dev. (2014)	Mean (2016)	Std. Dev. (2016)
Relative Health Deprivation	Row Kakwani individual health deprivation index; value range: [0, 1]	0.252	0.184	0.244	0.170
Elderly care services	See Table 1	0.069	0.253	0.067	0.251
	Pension insurance	0.807	0.394	1 *	0 *
	Senior benefits	0.276	0.447	0.325	0.468
	Financial support	3.807	2.106	3.965	1.922
	Living support	2.358	1.476	2.742	1.401
	Emotional support	3.326	1.824	3.612	1.670
Age	Continuous variable	71	8.071	70	7.596
Gender	1 = female, 0 = male	0.534	0.499	0.502	0.500
Marital status	1 = married, 0 = single	0.641	0.480	0.705	0.456
Place of residence	1 = rural, 0 = urban	0.423	0.494	0.455	0.498
Number of people living together	Continuous variable	3.250	1.893	2.759	1.307
Number of living children		3.098	1.567	2.716	1.399
Number of chronic diseases	Continuous variable	1.950	1.812	1.364	1.646
Income in past year (ln)	Continuous variable	9.004	1.329	9.055	1.347

* Note: in the 2016 data, the sample reached full pension coverage.

presents the descriptive statistics of the variables, showing a slight decrease in the relative deprivation index from 0.252 in 2014 to 0.244 in 2016. In the social security dimension, elderly care service utilization was less than 1%, with the number of users in 2016 (0.067) nearly matching that in 2014 (0.069). In 2016, pension coverage achieved full sample coverage; the proportion of elderly receiving preferential treatment rose from 28% to 33%. In terms of family support, the elderly received increased levels of economic (3.965), emotional (3.612), and living support (2.742) in 2016 compared with 2014 levels (3.807, 3.326, and 2.358, respectively). Regarding control variables, the average age was around 70, and there were slightly more female than male elderly persons. The partnered elderly (around 71%) outnumbered the single elderly (64%), and the rural elderly outnumbered the urban–rural elderly. Compared with 2014, in 2016, the number of cohabitants decreased by 50%, and the number of living children and chronic diseases were both lower. As can be seen from

Table 3. Sample elderly relative health deprivation status.

Variable Name	Variable Assignment	Relative Health Deprivation *
		2014	2016
Elderly care services	Used	0.253	0.240
	Not used	0.252	0.244
Pension insurance	Received	0.251	0.244
	Not received	0.257	*
Senior benefits	Enjoyed	0.228	0.225
	Not enjoyed	0.262	0.253
Financial support	Received	0.248	0.243
	Not received	0.268	0.250
Living support	Received	0.250	0.242
	Not received	0.256	0.248
Emotional support	Received	0.253	0.238
	Not received	0.251	0.266
Total of all samples		0.252	0.244

* Note: relative health deprivation, measured as the average relative deprivation within subgroups; in the 2016 data, the sample reached full pension coverage.

, the relative deprivation index for the health of the elderly who enjoyed social security benefits and received family support showed a downward trend.

4. Analysis of Empirical Results

4.1. Estimation of Impact of Social Elderly Care Security on Elderly Relative Health Deprivation

The baseline regression results from Model 2 in

Table 4. Panel data Tobit model baseline regression results (dependent variable: relative health deprivation).

Variable Name	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Age	0.0003 (0.0001)	0.0004 ** (0.0002)	0.0001 (0.0002)	0.0003 (0.0002)	0.0003 (0.0002)	−0.245 ** (0.282)	controlled
Gender	0.004 (0.003)	0.004 (0.003)	0.004 (0.003)	0.004 (0.003)	0.005 * (0.003)	−2.173 *** (0.147)
Marital status	−0.008 ** (0.003)	−0.007 ** (0.003)	−0.006 ** (0.003)	−0.006 ** (0.003)	−0.006 ** (0.003)	0.793 *** (0.168)
Place of residence	0.019 *** (0.003)	0.016 *** (0.003)	0.017 *** (0.003)	0.014 *** (0.003)	0.015 *** (0.003)	−1.658 *** (0.166)
Number of cohabitants	−0.001 (0.001)	−0.001 (0.001)	−0.002 (0.001)	−0.002 (0.001)	−0.002 (0.001)	0.018 (0.045)
Number of living children	0.004 *** (0.001)	0.003 *** (0.001)	0.003 *** (0.001)	0.003 *** (0.001)	0.003 *** (0.001)	−0.293 *** (0.055)
Number of chronic diseases	0.029 *** (0.001)	0.029 *** (0.001)	0.029 *** (0.001)	0.029 *** (0.001)	0.028 *** (0.001)	−0.181 *** (0.041)
Ln income in past year	−0.012 *** (0.001)	−0.011 *** (0.001)	−0.011 *** (0.001)	−0.010 *** (0.001)	−0.011 *** (0.019)	1.029 *** (0.062)
Elderly care services		−0.004 (0.005)		−0.003 (0.005)	−0.003 (0.005)	0.746 ** (0.282)
Pension insurance		0.004 (0.005)		0.005 (0.005)	0.011 ** (0.005)	1.322 *** (0.257)
Senior benefits		−0.012 *** (0.003)		−0.010 ** (0.003)	−0.009 ** (0.003)	1.395 *** (0.176)
Financial support			−0.002 ** (0.001)	−0.002 ** (0.001)	−0.002 ** (0.001)	0.312 *** (0.037)
Emotional support			−0.004 *** (0.001)	−0.004 *** (0.001)	−0.004 *** (0.001)	0.509 *** (0.043)
Living support			−0.001 (0.001)	−0.001 (0.001)	−0.001 (0.001)	0.054 (0.051)
Individual attitudes towards aging						−0.001 *** (0.0001)
Time-fixed effects	Yes	Yes	Yes	No	Yes	Yes
Sample size	16,327	16,327	16,295	16,295	16,295	16,327

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with standard errors in parentheses.

show that after controlling for demographic, economic, and social characteristics, as well as time-fixed effects, among the measures of social elderly care security (utilization of elderly care services, entitlement to pension insurance, and senior benefits), only senior benefits had a mitigating effect on relative health deprivation among the elderly. The elderly who received preferential treatment exhibited a 1.2% reduction in the level of relative health deprivation compared with the control group, and this effect was statistically significant at the 1% confidence level. Additionally, the use of elderly care services was associated with a reduction in elderly people’s perception of relative health deprivation, while pension insurance was linked to an increase in it; however, these associations did not reach statistical significance. This finding indirectly implies that in contrast to direct financial support and paid elderly care services, the provision of diverse forms of social preferential treatment may more effectively alleviate the sense of relative health deprivation among the elderly.

4.2. Estimation of Impact of Family Support on Elderly Relative Health Deprivation

Controlling for demographic, socio-economic, and societal variables, as well as time-fixed effects, Model 3 in Table 4 demonstrates the impact of familial support, quantified by financial support, living support, and emotional support, on the relative health deprivation experienced by the elderly. The regression analysis suggests that familial support significantly alleviates elderly people’s perception of relative health deprivation. Those who received emotional support exhibited a 0.4% reduction in the level of relative health deprivation compared with the control group, and this effect was statistically significant at the 1% confidence level. Elderly persons receiving financial support showed a 0.2% decrease in relative health deprivation levels compared with the control group, and this difference was statistically significant at the 5% confidence level.

4.3. Estimation of Joint Impact of Social and Family Support on Elderly Relative Health Deprivation

Model 4 in Table 4 presents the regression analysis examining the joint impact of social security and familial support on health deprivation among the elderly. After accounting for time-fixed effects and variables related to demographics, society, and the economy, the model results indicate that elderly individuals who benefited from preferential elderly treatment (senior benefits) ($\beta$ = −0.010, p < 0.05) and received financial support ($\beta$ = −0.002, p < 0.05) and emotional support ($\beta$ = −0.004, p < 0.01) experienced a reduced level of health deprivation compared with those without such benefits. Having pension insurance increased the sense of relative health deprivation among the elderly, while utilizing elderly care services and receiving care giving lowered it, but these effects were not statistically significant. Hence, Hypothesis 1 was partially supported: elderly individuals with access to preferential treatment and who receive familial financial and emotional support report lower levels of relative health deprivation.

4.4. Estimation of Combined Effect of Social Security for the Elderly and Family Support on Individuals’ Attitudes toward Aging

After adjusting for time-fixed effects and socio-economic demographic variables, the estimates from Model 6 in Table 4 show that older individuals who utilized elderly care services ($\beta$ = 0.746, p < 5%), had access to pension insurance ($\beta$ = 1.322, p < 1%), received preferential treatment as senior citizens (senior benefits) ($\beta$ = 1.395, p < 1%), and received financial ($\beta$ = 0.312, p < 1%) and emotional support ($\beta$ = 0.509, p < 5%) exhibited a more positive attitude towards aging. However, the impact of family care giving on this attitude was not significant. Hypothesis 2 was thus confirmed, indicating that social security and family support are associated with a more positive attitude towards aging among older individuals.

4.5. Estimation of Combined Effect of Individual Attitudes on Relative Health Deprivation in the Elderly

After controlling for time-fixed effects and socio-economic demographic variables in Model 7 (Table 4), it was found that a more positive attitude towards aging ($\beta$ = −0.001, p < 1%) was associated with a lower sense of relative deprivation in elderly health. The third hypothesis was supported.

4.6. Analysis of Impact of Control Variables on Elderly Relative Health Deprivation

The estimation results of the full regression model (Model 5) indicate that the marital status, the number of cohabitants, and the income in the past year had a significant negative impact on relative health deprivation among the elderly. In contrast, the place of residence, the number of living children, and the number of chronic diseases had a significant positive impact. Specifically, an increase in income, the number of cohabitants, and having a spouse significantly reduced the level of health deprivation among the elderly. Conversely, living in rural areas, having more living children, and a higher number of chronic diseases increased the level of health deprivation. These findings imply that for the elderly, the companionship and care of a spouse, along with a certain income level, may be more reliable than having multiple children.

5. Robustness Tests

5.1. Ridge Regression

Traditional regression analysis often suffers from multicollinearity, which may cause the signs of the regression coefficients to be opposite to what is actually the case, making the independent variables insignificant or less significant. Step-wise regression, by technically removing variables with multicollinearity, may hinder the study of the intended variables. In contrast, ridge regression exhibits significantly higher tolerance for multicollinearity than ordinary least squares regression, offering a robust alternative for dealing with this issue (

Table 5. Ridge regression.

Variable Name	Relative Health Deprivation
Control variables	controlled
Elderly care services	0.008 (1.218)
Pension insurance	0.005 (0.809)
Senior benefits	−0.015 ** (−3.616)
Financial support	−0.002 * (−1.824)
Emotional support	−0.005 ** (−4.869)
Living support	0.001 (1.018)
Sample size	16,295
$R^2$	0.053

Note: ** and * indicate significance at the 5% and 10% levels, respectively, with standard errors in parentheses.

).

Enjoying senior benefits, financial support, and emotional support can significantly reduce feelings of health deprivation among the elderly, consistent with the Tobit regression results in Table 4, indicating the robustness of the regression.

5.2. Propensity Score Matching (PSM)

This study’s earlier findings indicate that senior benefits, living support, and emotional support can mitigate the relative sense of health deprivation among the elderly. However, besides endogeneity issues arising from reciprocal causation, the selection process for providing these services is also influenced by various factors and is not random. For instance, the elderly with higher vitality and those living in urban areas may enjoy more senior benefits, while those in poor health may be more likely to receive care giving, feeling a stronger sense of relative health deprivation. To tackle the endogeneity issues stemming from self-selection bias, in this study, we utilized three common matching methods to further analyze the net impact of related variables on relative health deprivation among the elderly (

Table 6. The impact of preferential treatment for the elderly on relative health deprivation according to different matching methods.

Matching Method		Treatment Group	Control Group	ATT	Standard Error	T-Value
Nearest-neighbor matching method	Before matching	0.133	0.173	−0.039	0.003	−12.59
	After matching	0.133	0.148	−0.015	0.005	−2.89
Radius matching method	Before matching	0.133	0.173	−0.039	0.003	−12.58
	After matching	0.133	0.145	−0.012	0.004	−4.96
Kernel matching method	Before matching	0.133	0.173	−0.039	0.003	−12.58
	After matching	0.133	0.145	−0.012	0.004	−3.19

,

Table 7. The effect of emotional comfort on relative health deprivation according to different matching methods.

Matching Method		Treatment Group	Control Group	ATT	Standard Error	T-Value
Nearest-neighbor matching method	Before matching	0.156	0.228	−0.072	0.006	−11.54
	After matching	0.158	0.205	−0.047	0.010	−4.66
Radius matching method	Before matching	0.156	0.228	−0.072	0.006	−11.54
	After matching	0.158	0.213	−0.055	0.008	−6.88
Kernel matching method	Before matching	0.156	0.228	−0.072	0.006	−11.54
	After matching	0.158	0.220	−0.063	0.007	−8.50

and

Table 8. Heterogeneity regression results.

Matching Method		Treatment Group	Control Group	ATT	Standard Error	T-Value
Nearest-neighbor matching method	Before matching	0.157	0.178	−0.021	0.004	−5.57
	After matching	0.157	0.178	−0.021	0.005	−4.39
Radius matching method	Before matching	0.157	0.178	−0.021	0.004	−5.57
	After matching	0.157	0.180	−0.023	0.004	−6.01
Kernel matching method	Before matching	0.157	0.178	−0.021	0.004	−5.57
	After matching	0.157	0.178	−0.021	0.004	−5.51

).

The results from the table show that when applying different matching methods, the estimated Average Treatment Effect on the Treated (ATT) is significant at the 1% level. This indicates that after accounting for sample selection bias, enjoying senior benefits and receiving financial support and emotional support can significantly reduce the sense of health deprivation among the elderly. This is consistent with the estimates from the baseline regression model, suggesting that the regression results are robust.

6. Heterogeneity Analysis and Mechanism Testing

6.1. Heterogeneity Test

The analysis by gender indicated that financial support and emotional support inhibited the level of relative health deprivation among the elderly regardless of gender. However, within the dimension of social security for the elderly, only participation in pension insurance was associated with a significant increase in relative health deprivation among elderly men, while for elderly women, only enjoying senior benefits significantly reduced the relative health deprivation level (

Table 9. Heterogeneity regression results.

Variable Name	Dependent Variable: Relative Health Deprivation
	Urban	Rural	Male	Female	Young Age (60–69 Years)	Middle Age (70–79 Years)	Older Age (80+ Years)
Elderly care services	0.0001 (0.006)	−0.006 (0.010)	−0.001 (0.008)	−0.004 (0.007)	0.002 (0.008)	−0.015 (0.009)	0.010 (0.013)
Pension insurance	−0.004 (0.007)	0.012 * (0.007)	0.016 ** (0.008)	−0.003 (0.006)	−0.006 (0.006)	0.022 ** (0.009)	−0.0004 (0.013)
Senior benefits	−0.005 (0.004)	−0.026 *** (0.007)	−0.006 (0.005)	−0.014 ** (0.005)	−0.002 (0.005)	−0.026 *** (0.006)	−0.007 (0.009)
Financial support	−0.002 ** (0.001)	−0.003 ** (0.001)	−0.002 ** (0.001)	−0.002 ** (0.001)	−0.002 ** (0.001)	−0.003 ** (0.001)	−0.001 (0.002)
Living support	0.0003 (0.001)	−0.002 (0.002)	−0.002 (0.001)	0.001 (0.001)	−0.002 (0.001)	−0.002 (0.002)	0.004 (0.003)
Emotional support	−0.004 *** (0.001)	−0.004 ** (0.001)	−0.005 *** (0.001)	−0.003 ** (0.001)	−0.005 *** (0.001)	−0.002 (0.001)	−0.004 ** (0.002)
Control variables Time-fixed effects	Controlled
Sample size	9131	7164	7857	8438	8428	5374	2493

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with standard errors in parentheses.

).

The analysis by residential area indicated that within the dimension of family support for the elderly, increased financial support and emotional support significantly inhibited the sense of relative health deprivation among the elderly, both in urban and rural areas. However, within the dimension of social security for the elderly, the effects of the three variables on the urban elderly were no longer significant. In contrast, in rural areas, senior benefits significantly inhibited the sense of relative health deprivation among the elderly, while participation in pension insurance increased it.

Age-based testing revealed that the dimension of social security for the elderly did not affect the level of relative health deprivation among elderly persons aged 60–69. For the elderly aged 70–79, changes began to appear: the use of elderly services (p = 0.1, indicating a marginally significant effect) and enjoying senior benefits reduced their relative health deprivation, while participation in pension insurance increased it. For the oldest elderly group, emotional support reduced their sense of health deprivation, while living support (p = 0.1, i.e., marginally significant) increased it.

6.2. Mechanism Testing

Since the results of the Sobel test and Bootstrap test were nearly identical, here, we only report the results of the former, which show that the mediating effect of individuals’ attitudes toward aging accounted for 15.75% of the total impact of senior benefits on relative health deprivation (p < 0.01), 17.61% of the impact of financial support (p < 0.01), and 15.92% of the impact of emotional support (p < 0.01). Enjoying senior benefits and receiving financial support and emotional support from children positively influenced the attitudes of the elderly towards aging, thereby reducing their sense of relative health deprivation. Hypothesis 4 was partially verified (for details, see

Table 10. Heterogeneity regression results.

Core Independent Variable	Mediator Variable	Effect	Effect Coefficient	Sobel Test	Goodman Test		Proportion of Mediation Effect
					Test 1	Test 2
Senior benefits	Individual attitudes towards aging	Indirect effect	−0.002 ***(z = 6.953)	−0.002 ***(z = 6.954)	−0.002 ***(z = 6.936)	−0.002 ***(z = 6.971)	15.75%
		Direct effect	−0.013 *** (z = 4.097)
Financial support		Indirect effect	−0.001 ***(z = 6.844)	−0.0005 ***(z = 6.844)	−0.0005 ***(z = 6.826)	−0.0005 ***(z = 6.862)	17.61%
		Direct effect	−0.002 ***(z = 3.444)
Emotional support		Indirect effect	−0.001 ***(z = 7.435)	−0.001 ***(z = 7.435)	−0.001 ***(z = 7.421)	−0.001 ***(z = 7.45)	15.92%
		Direct effect	−0.004 ***(z = 4.935)

Note: *** indicates significance at the 1%, with standard errors in parentheses.

and Figure 2).

7. Discussion and Research Conclusions

In this study, based on data from the China Health and Retirement Longitudinal Study, we measured the health inequality among the Chinese elderly by using the Kakwani index of relative deprivation. We constructed an empirical model to systematically examine the impact mechanisms and heterogeneity of different elderly support systems on health inequality. Below, we report our findings and the conclusions drawn from them.

Firstly, access to elderly social security and receiving familial support significantly reduce elderly people’s sense of relative health deprivation in China. In our study, within the social security dimension, only the provision of senior benefits significantly suppressed relative health deprivation among the elderly, while participation in pension insurance initially increased this sense of deprivation. However, its effect became non-significant after accounting for temporal dynamics, which contradicts the research findings obtained by Been, J. et al. (2017) [20]. First, the research findings may be influenced by sample selection, as different databases might lead to variations in the conclusions. Second, the time span of data collection is a critical factor: self-administered questionnaires allow for more flexible timing, potentially revealing more pronounced temporal effects. Lastly, in this study, we employed a simplified approach to variable setting by using participation in pension insurance as a measure of its impact on health inequality, without delving into the specific effects of pension amounts on health disparities. Regarding family support, financial and emotional backing from children notably lowered elderly people’s sense of relative health deprivation, while daily care did not affect it. This is in line with the observations by Graffigna, G. et al. (2021) [13] on the general weakening of family service functions.

Secondly, enjoying social security benefits for the elderly and receiving family support can foster a more positive attitude towards aging among the elderly. Specifically, older individuals who used elderly services, had pension insurance, enjoyed senior benefits, and received financial and emotional support exhibited a more positive attitude towards aging. This finding partially aligns with Wang Liu Y.’s (2023) [22] belief that emotional comfort has the most significant impact on elderly people’s self-perception of aging, followed by living and financial support, and is consistent with Smith J.’s (2022) [25] observation that social security and children’s financial support can lead to positive changes in attitudes towards aging.

Third, the more positive an individual’s attitude towards aging, the lower their sense of relative health deprivation, with the former serving as a mediator between different types of elderly support and relative health deprivation. Enjoying senior benefits and receiving financial support and emotional comfort from children positively influenced the attitudes of the elderly towards aging, thereby reducing their sense of relative health deprivation.

Fourth, based on heterogeneity tests, we found the following: The inhibitory effects of financial support and emotional support on relative health deprivation among the elderly did not differ by gender. However, participating in pension insurance significantly increased relative health deprivation among elderly men, while elderly women showed a significantly lower level of health deprivation when enjoying senior benefits. Regardless of living in urban or rural areas, an increase in financial and emotional support significantly inhibited the sense of relative health deprivation among the elderly. In the dimension of social security for the elderly, the three variables no longer significantly affected the urban elderly but did so in rural areas, where the availability of senior benefits is significantly reduced, and pension insurance increased the sense of health deprivation. Social security for the elderly did not affect the younger elderly but started to show changes for the middle-aged elderly, among whom using elderly services and enjoying senior benefits are less common, and pension insurance increased their health deprivation level. For the older elderly, emotional support reduced their sense of health deprivation, while living support increased it. This is consistent with previous findings (Been, J. et al., 2017) [20] indicating that pension increases have widened the health gap among middle-aged and younger elderly populations. In China, the pension contribution levels for rural elderly individuals are generally low, with the vast majority falling into the category of low pension income, which aligns with the current situation.

Although this study’s findings offer preliminary insights into the relationship between senior care and health inequality among the elderly, several limitations and areas for further exploration remain: (1) While the sample partially represents the Chinese elderly population, further verification is needed regarding its representativeness and the implications for the broader applicability of the findings. Moreover, despite efforts to control for temporal effects, their long-term impact on the findings deserves deeper investigation. Notably, the influence of pension insurance on the sense of health deprivation among the elderly is no longer statistically significant after controlling for temporal effects, indicating a need for further research into both its short-term and long-term impacts. (2) This study’s results indicate that economic and emotional support from children can effectively alleviate the sense of health deprivation among the elderly, whereas the effect of living support is less impactful. Other forms of familial elderly care support and their health implications for the elderly necessitate more in-depth research. Additionally, the mediating function of attitudes toward aging between elderly care support and the sense of health deprivation warrants further clarification in terms of its underlying mechanisms and extent. (3) Long-term follow-up studies are especially crucial for a comprehensive understanding of the long-term effects of senior care support on health inequality among the elderly. Given that cultural disparities may influence both senior care support and health inequality, cross-cultural comparative research could offer a more expansive viewpoint.

Health inequality is one of the key indicators for measuring the level of sustainable social development, and its existence and persistence have various negative impacts on sustainability.Firstly, at the socio-economic level, health inequality can lead to imbalances in the labor market, affecting economic growth and social productivity. In particular, health disparities among the elderly may limit their re-employment opportunities and work capacity, thereby impacting economic development and social welfare. Secondly, health inequality increases the public health burden. Uneven demand for medical resources may lead to inefficient resource allocation, exacerbating the financial strain on the healthcare system and potentially leading to a decline in the quality of healthcare services. Additionally, health inequality reflects the inequitable distribution of social resources, which may lead to social discontent and tension, affecting social stability and harmony. Especially in a populous country like China, health inequality may exacerbate social stratification between urban and rural areas, genders, and age groups, affecting social equity and justice. In terms of environmental sustainability, health inequality is closely related to environmental factors. Poor living and working conditions are often associated with higher health risks. This not only affects individual health but also poses a threat to the long-term sustainability of the environment. Lastly, in terms of intergenerational effects, health inequality may indirectly impact the growth environment and health status of the next generation by affecting the quality of life and mental health of the elderly, increasing the family burden, and affecting children’s education and development.

Therefore, addressing health inequality requires not only improving social security and pension services but also considering the feedback effects of policies and systems to achieve fairness and effectiveness in policy implementation, thereby supporting overall sustainable social development.

Policy Implications:

(1)

Refine the social pension insurance system. The social pension insurance system is crucial to ensuring the basic livelihood and health of the elderly and is a key factor in bolstering their self-esteem and confidence. Continued investment and reform in elderly care services and preferential treatment for seniors are essential to expanding coverage, enhancing benefit levels, and improving the system’s sustainability and equity. This will meet the diverse needs of the elderly population, promoting economic sustainability by reducing the strain on labor markets and social welfare systems.

(2)

Reinforce family support in elderly care. Support from family members is a vital aspect of traditional Chinese culture and an effective means to reduce health inequalities among the elderly. Promoting a culture of filial piety and encouraging children to provide economic and emotional support to their parents will foster family harmony and closeness. This not only respects the elderly’s autonomy but also safeguards their dignity and rights, contributing to social stability and harmony, which are cornerstones of sustainable societies.

(3)

Foster a positive attitude towards aging. An individual’s attitude towards aging significantly mediates health inequalities among the elderly and is pivotal to their sense of well-being. Education and guidance for the elderly, including community health campaigns and mental health services, should be strengthened to help them develop a positive view of aging and recognize the value of life in their later years. This approach encourages active social engagement among the elderly, leveraging their strengths and potential, fulfilling their self-worth, and enhancing their societal contributions, which are essential for intergenerational equity and sustainability.

(4)

Make health policies that account for diversity among the elderly. Given the varying health statuses and needs of the elderly based on gender, residence, and age group, targeted measures should be adopted in health policy and resource allocation. There should be an increased focus on health investments for women, rural seniors, and the oldest old to bridge the gap in health resource accessibility. These policies will address the environmental and social determinants of health, ensuring that all segments of the elderly population can contribute to and benefit from sustainable development.