Living Long and Well: Cross-Temporal Meta-Analytic Evidence on Elderly Chinese Health-Related Quality of Life

Abstract

China has been successful in getting her people to live longer. But, merely adding years to life and not life to years poses immense socio-economic challenges. However, not much is known about the effects of government policy and program interventions on changes in how well the elderly live over the years. Accordingly, we cross-temporally meta-analyzed 45 research reports (N = 36,352) that utilized the health-related quality of life (HRQoL) scale (SF-36) from 2000 to 2020. We found that: (1) the bodily pain, general health, vitality, and mental health of the elderly deteriorated over time; however, their physical and emotional roles, as well as their social functioning, improved with time; (2) the rising dependency ratio impoverished the HRQoL of the elderly; (3) the HRQoL indicators of the elderly revealed positive gains under the home-based care model whilst they showed a downward trend under the institutional pension model; (4) the HRQoL indicators of the elderly in economically developed areas produced mixed results; but, they all worsened over the years in economically underdeveloped areas. Thus, more investment efforts from the government and private entities are needed to reduce the dependency ratio and to improve the lives of the elderly under institutional care and/or in economically underdeveloped areas.

1. Introduction

The people of China now live longer. The increased life expectancy is driven by healthy living, the elimination of infectious pathogens, and the treatment of non-communicable diseases [1,2]. These developments have been possible due to a colossal investment by the Chinese government into food security, medical advancement, and making available more affordable and accessible healthcare [3,4]. However, some of the policies of the government directed at improving the lives of Chinese people have induced structural population shifts [5], with epidemiological and socio-economic implications [1]. For instance, the spillover effects of the one-child policy are the looming low-fertility trap and rising dependency ratio (which reflects the number of elderly people per one hundred working-age adults) [6]. Another is that the rapid economic development in urban centers has sped up rural–urban migration, creating left-behind children and elderly people with little support systems [1,6,7]. At present, there are 267.36 million Chinese people aged 60 years and above, accounting for 18.9% of the population compared to 13.9% of the total population in 2010—just a little over a decade ago [8]. More troubling is the fact that about 200.56 million of the aging population are at least 65 years old—the largest globally. The health of the elderly (aged 60 and above) in China is, therefore, a matter of national security with global implications that needs to be addressed.

Meeting the healthcare needs of the elderly is dependent on an intricate mix of fine insights into the physical and psychological health of the elderly, their socio-economic capital, and other critical support systems, including healthcare infrastructure. As China officially joins the “aging nations”, it is crucial to recognize her unique characteristics, such as getting old before getting rich, large regional wealth differences, family dependency, and, importantly, “being not ready to get old” [6,9]. How well the elderly live in China is a key question of scientific inquiry. Most often, the elderly’s health-related quality of life (HRQoL)—defined as “a multidimensional concept with both objective and subjective factors that refer to general satisfaction with life or its components” [10]—has been used as an indicator of a good/bad life [11,12,13,14].

Findings from extant studies depict an inconsistent picture of the elderly’s HRQoL. Some studies show a rising HRQoL of the elderly (e.g., [15]) whilst others indicate a falling trend (e.g., [16]). Additionally, the HRQoL of the elderly in underdeveloped areas [17,18], and those in institutional care [19,20], is found to be poor. With rapid economic development lifting millions from rural poverty through “digital dividend” [21] and consistent government investment in institutional care [3,4,22], the nature of the evolution of the HRQoL of the elderly across care models and regions of domicile over the years is uncertain. Additionally, the rising dependency ratio implies that little time and attention may be dedicated to caring for the elderly [23], particularly those in the home-based care model. There is therefore an urgent need to establish the patterns of change in the elderly’s HRQoL through time to put investment in intervention policies and programs [3] geared toward improving the HRQoL of the elderly into perspective. Our research responds to this need. We therefore attempt to answer the following questions: (1) Is the HRQoL of the elderly in China rising or falling over the years (i.e., from 2000 to 2020)?; (2) To what extent does this change (rise or fall) differ across the different elderly care models adopted and the regions of domicile?; (3) Lastly, how does the rising dependency ratio impact the HRQoL of the elderly over the years?

To answer the aforementioned questions, we adopt a cross-temporal meta-analytic (CTMA) approach to examine the changing trends in the HRQoL of the elderly in China from 2000 to 2020, with age (publication year) as the independent variable. This approach is considered appropriate as traditional meta-analytic methods fail to account for the “age effect” of perceptual variables [24]. That is, the CTMA technique eliminates measurement errors introduced by age effect (publication year) and can, thus, disentangle the nature of association between psychological variables and socio-economic indicators to explain how social and economic changes affect individual psychological development [25]. The dynamic nature of the current socio-cultural and techno-economic developments necessitates continual evaluation of how these developments influence important life goals and psychological wellness [26,27,28]. Therefore, CTMA has been utilized to reveal changes in social support systems for the elderly [29] and the mental health of middle school students in China [24,28]. In addition to the within-scale variability assessment, we employ conventional meta-analytic techniques to examine the differences in HRQoL across groups of elderly people.

2. Materials and Methods

2.1. Research Tool: The Medical Outcomes Study (MOS) 36-Item Short-Form Health Survey (SF-36)

The MOS 36-Item Short-Form Health Survey (SF-36), which originated from the work of the Santa Monica Rand Company in the 1970s, has been improved and used in health research [30]. The SF-36 is a self-rating scale with 36 items grouped into 2 main components—the Physical Component Summary (PCS) and the Mental Component Summary (MCS), with 4 subscales each. The PCS’s subscales include Physical Functioning (PF: whether or not an individual’s health status hinders normal physiological activities); Role-Physical (RP: the functional limitation caused by physiological health problems); Bodily Pain (BP: the degree of pain and the influence of pain on daily activities); and General Health (GH: an individual’s evaluation of his/her own health status and development trend). The MCS encompasses Vitality (VT: an individual’s subjective feelings about his/her own energy and fatigue); Social Functioning (SF: the influence of physiological and psychological problems on the quantity and quality of social activities); Role-Emotional (RE: functional limitation caused by emotional problems); and Mental Health (MH: consists of four dimensions—motivation, depression, behavior or emotions being out of control, and psychological subjective feelings) [30]. The SF-36 scale is widely used to measure the HRQoL of the elderly in China [31]. The scores of the scale are all converted into percentages [32] and, therefore, range from 0 to 100. Higher scores reflect a better HRQoL.

2.2. Literature Search Strategy

We adopted the following search criteria in retrieving documents: (1) the paper must be published at home or abroad; (2) the research subjects were sampled from the elderly population in mainland China, excluding the elderly in China’s Special Administrative Regions, such as Hong Kong, Macao, and Taiwan; (3) the SF-36 scale is the tool used to investigate the HRQoL of the elderly; (4) complete data, including the sample size, average, and standard deviation of each dimension of the scale; and (5) using the latest and most complete documents in cases of duplicate publications.

2.3. Document Selection and Coding

To retrieve a comprehensive body of work, we searched famed databases, including the CNKI database, Wanfang database, VIP database, and other Chinese and foreign databases, such as Elsevier, ProQuest, and Wiley. Our search terms were quality of life, the elderly, health-related quality of life, and the elderly and we searched for documents up to January 2022. The search returned 2659 articles from the HowNet database; 23,719 articles from the Wanfang database; 1345 articles from the VIP database; 17 articles from the Wiley database; and 1337 articles from the SpringerLink database. After reading the title and abstract, the irrelevant documents were excluded, leaving 371 articles. After reading the full text—screening out papers with incomplete data, inconsistent uses of scales, and other unqualified documents—45 articles were finally retained (see Figure 1). The years of data collection of the studies included in this paper (hereinafter referred to as “years”) were obtained by subtracting 2 years from the year of publication in cases where the date of data collection is not explicitly stated. Therefore, the age range of this study is from 2000 to 2020.

The 45 selected documents were carefully coded. For studies that utilized sub-data rather than the total data, the results of such sub-studies were weighted and synthesized, according to the following two formulas:

$$\overline{x}={\displaystyle \sum x_i}{n}_i/{\displaystyle \sum n_i}$$

$$S_T=\sqrt{\left[{\displaystyle \sum n_i{s}_{{}_i}^2+{\displaystyle \sum n_i{(x_i-\overline{x_i})}^2}}\right]/{\displaystyle \sum n_i}}$$

where x, S_T, n_i, x_i, and S_i, respectively, represent a composite average; the resultant standard deviation; the sample size of a study; the average; and the standard deviation.

To calculate the magnitude of the changes in the HRQoL scores over time, we used the regression equation: y = Bx + C, where X = the year, Y = the predicted mean HRQoL scores, B = the non-standardized regression coefficient, and C = the constant term. This equation yielded the expected average HRQoL score for particular years.

In accordance with the demands of CTMA, three steps were followed. First, each document was given a unique number and the basic data of the document, such as sample size (N), average value (M), and standard deviation (SD), as well as the publication year and the data collection year. Second, these documents were entered into the database. Third, the reported results for a subpopulation of the elderly, including their living areas, were also coded (see

Table 1. Cross-Temporal Variable Coding Table.

Variable	Coding	Literature Quantity
endowment pattern	1 = home-based care model	14
	2 = institutional pension model	13
area	1 = economically developed areas	11
	2 = economically underdeveloped areas	34

for specific coding themes).

2.4. Analytic Strategy

We utilized Microsoft Excel 2010 and IBM SPSS version 22 to process and analyze the valid data. The formula for the calculation of the effect d is d = (x₂₀₀₄ − x₂₀₂₀)/SD. We derived our coefficient of determination using this function: r² = d²/(d² + 16). The difference was considered statistically significant at p < 0.05.

3. Results

3.1. The Overall Change of the HRQoL of the Elderly in China over the Years

To establish the overall patterns of change in terms of the indicators of the HRQoL of the elderly over the years, we generated scatter charts (example below) that show the relations between the factors of the SF-36 scale and the publication year or age (that is, the document’s published year minus 2 years). The estimated results show that the linear model fits well (see Figure 2).

In studies using CTMA, the sample size of each document interferes with the psychometric-chronological results. Therefore, while the sample size was weighted, the age (that is, the document’s published year minus 2 years) was used as the independent variable and the mean of each factor of the SF-36 was used as the dependent variable for the correlation analyses. Age correlated significantly negatively with bodily pain, general health, vitality, and mental health whilst correlating significantly positively with role-physical and role-emotional and social functioning. Physical functioning was not significantly correlated with age (see

Table 2. Correlations between factors of the HRQoL of the elderly and age (published year minus 2 years).

	Uncontrolled Sample Size		After Controlling the Sample Size
Index	r	r2	β	β2
PF	−0.294	0.086	0.007	0.000049
RP	−0.185	0.034	0.108 **	0.011 **
BP	−0.176	0.031	−0.063 **	0.004 **
GH	−0.196	0.038	−0.175 **	0.031 **
VT	−0.284	0.081	−0.264 **	0.07 **
SF	0.004	0.000016	0.092 **	0.008 **
RE	−0.233	0.054	0.086 **	0.007 **
MH	−0.226	51	−0.136 **	0.018 **

Note: ** p < 0.01. r = correlation coefficient of the uncontrolled sample size, β = correlation coefficient after the control sample size. Applies across tables.

).

In order to establish the quantum of change, we performed regression analysis using the standard deviation of each study [33]. First, using the regression equation weighted by samples, the change in the mean value for each factor of HRQoL reported in studies from 2000 to 2020 was calculated. The regression equation of this study was y = Bx + C, where B represents the non-standardized regression coefficient, X is the year, C is the constant term, and Y is the average. The average standard deviation was obtained by averaging the standard deviations of all studies. This method of using individual-level variables effectively avoids ecological fallacies [34]. In this study, following best practices [28,35], the effect (d) and the interpretation rate (r²) were calculated.

It can be seen from

Table 3. Changes in the HRQoL of the elderly from 2000 to 2020.

Index	M2002	M2020	ΔM	SD	d	r2
PF	69.44	69.76	0.32	20.6	0.02	0.0001
RP	609.44	614.62	5.18	31.73	0.16	0.01
BP	71.82	69.07	−2.75	18.45	−0.15	0.01
GH	57.56	52.03	−5.53	16.62	−0.33	0.03
VT	72.00	64.75	−7.25	15.26	−0.48	0.05
SF	73.06	75.76	2.7	20.01	0.13	0.004
RE	65.79	68.89	3.1	32.54	0.10	0.002
MH	71.59	67.41	−4.18	14.71	−0.28	0.02

Note: ΔM = M₂₀₂₀ − M₂₀₀₂, d = ΔM/M_SD, r² = d²/(d² + 4), M_SD is the 18-year average standard deviation of each factor. Applies across tables.

that the total score of some indicators of the HRQoL of the elderly plummeted over time. Specifically, the elderly’s bodily pain, general health, vitality, and mental health got worse. As shown in Table 3, some indicators of the HRQoL of the elderly improved over the years. The particular indicators that got better over time include physical functioning, role-physical, social functioning, and role-emotional (please consult Table 3). Following standard effect size demarcation concerning an absolute value of an effect [36], we considered an effect size that was between 0.2 and 0.5 a small effect, between 0.5 and 0.8 a medium effect, and above 0.8 a big effect. It can be seen from Table 3 that general health, vitality, and mental health obtained small effects while other factors were close to small effects. Within 20 years, the average value of the eight factors of the HRQoL of the elderly ranged from 0.32 to 7.25, with an increase/decrease of about −0.48 to 0.16 standard deviations (that is, the d value in Table 3). This shows that although the HRQoL of the elderly has either increased or decreased, depending on which component of HRQoL one looks at, the rate of change is generally slow.

3.2. The Relationship between the Elderly Dependency Ratio and the HRQoL of the Elderly in China

To put the results from Table 3—where the elderly’s bodily pain, general health, vitality, and mental health were negatively correlated with age while role-physical and social functioning and role-emotional correlated positively with age—into a socio-economic perspective, we analyzed the effects of the dependency ratio on the HRQoL of the elderly. Using the dependency ratio of the elderly as a social indicator and an independent variable, we analyzed the HRQoL of the elderly. The first row of

Table 4. The old-age dependency ratios of the current year, five years prior, and ten years prior are correlated with the HRQoL of the elderly.

	PF	RP	BP	GH	VT	SF	RE	MH
The elderly dependency ratio of the current year	−0.158 **	−0.064 **	−0.145 **	−0.289 **	−0.366 **	−0.048 **	−0.009	−0.260 **
The elderly dependency ratio of five years ago	−0.178 **	−0.116 **	−0.128 **	−0.356 **	−0.395 **	−0.012 *	−0.059 **	−0.256 **
The elderly dependency ratio of ten years ago	−0.156 **	−0.079 **	−0.144 **	−0.298 **	−0.373 **	−0.046 **	−0.017 **	−0.267 **

Note: ** p < 0.01, * p < 0.5.

shows that the HRQoL of the elderly is negatively correlated with the dependency ratio of the elderly. That is, the rising dependency ratio impoverishes the HRQoL of the elderly. Accordingly, a high dependency ratio increases the burden of social and family support. In a typical family in China, a young couple needs to support two children and four elderly parents.

To explicate the direction of influence of the association between the dependency ratio of the elderly and the HRQoL, we performed a time-lag analysis [27]. The average HRQoL of the elderly was matched with the dependency ratio of the elderly five years ago and five years later and, then, the correlation was obtained. That is, the average value of each factor of HRQoL from 2000 to 2020 was correlated with the dependency ratio of the elderly in the 1995–2015 and 2005–2025 ranges, respectively. Since the dependency ratio data were from 1995 to 2021, the time-lag analysis after five years actually sought the correlation between the average values of various factors of HRQoL from 2000 to 2020 and the dependency ratio of the elderly from 2005 to 2020 for 15 years.

As shown in Table 4, the data of the dependency ratios of the elderly in the current year, that of five years earlier, and that of ten years earlier, after being weighted, are negatively correlated with various factors of HRQoL. Therefore, the dependency ratio of the elderly is an important factor affecting the HRQoL of the elderly. A significant negative correlation between the indicators of the elderly HRQoL and dependency was found in the data from five and ten years ago.

3.3. Changes in HRQoL of the Elderly in China under Different Care Models over the Years

Among the 45 articles utilized in this study, 14 articles were on home-based care and 13 articles were on institutional care. We performed a correlation analysis between the average scores of the indicators of the HRQoL of the elderly under different old-age care models with age. The results showed that the indicators of the HRQoL of the elderly mainly showed a positive relation with age in the home-based old-age care model. However, under the institutional pension model, all indicators of the HRQoL of the elderly were negatively correlated with age (please see

Table 5. Correlations between various factors of the HRQoL of the elderly under different care models and ages.

Endowment Pattern	PF	RP	BP	GH	VT	SF	RE	MH
home-based care model	0.161 **	0.090 **	−0.233 **	0.241 **	−0.025 **	0.085 **	0.273 **	−0.101 **
institutional pension model	−0.432 **	−0.690 **	−0.860 **	−0.709 **	−0.679 **	−0.380 **	−0.638 **	−0.664 **

Note: ** p < 0.01.

).

To obtain a better picture of the HRQoL of the elderly under different care models, we estimated the patterns of change over the years. As can be seen in

Table 6. The changes in the HRQoL of the elderly in China over the years under different care models.

Index	Home-Based Care Model				Institutional Pension Model
	ΔM	SD	d	r2	ΔM	SD	d	r2
PF	8.08	19.36	0.42	0.04	−22.91	21.53	−1.06	0.22
RP	5.38	29.94	0.18	0.008	−38.04	27.95	−1.36	0.32
BP	−11.81	17.04	−0.69	0.11	−48.92	17.61	−2.78	0.66
GH	5.85	14.45	0.40	0.04	−25.99	16.44	−1.58	0.38
VT	−0.77	14.06	−0.05	0.0006	−32.09	14.20	−2.26	0.56
SF	3.41	18.99	0.18	0.008	−18.92	17.90	−1.06	0.22
RE	9.25	28.48	0.32	0.02	−33.57	23.92	−1.40	0.33
MH	−3.91	14.38	−0.27	0.02	−36.54	14.69	−2.49	0.61

, each factor changes slowly in the home-based care model for the aged population (there is no large effect). Specifically, the effect size of the rate of decline in bodily pain is medium. The effect sizes of the changes in physical functioning, general health, role-emotional, and mental health are all small. Under the institutional care model, the scores of various factors of the HRQoL of the elderly dropped by 1.06–2.78 points. That is, there was a drop of 0.22–0.66 in standard deviation. According to the effect size demarcation standard [36], we found the changes in the indicators of HRQoL under the institutional care model for the aged population to be deteriorating rapidly (the decreasing range reaches a large effect).

3.4. Changes in the HRQoL of the Chinese Elderly in Different Regions over the Years

In accordance with the prior regional division of China in studying the HRQoL of the elderly on a regional basis [37], we explored whether the HRQoL of the elderly may be explainable by the region where they live.

Table 7. Correlation between various dimensions of the HRQoL of the elderly in different regions and ages.

	PF	RP	BP	GH	VT	SF	RE	MH
economically developed areas	0.338 **	0.135 **	−0.117 **	0.052 **	−0.075 **	0.042 **	−0.066 **	−0.053 **
economically underdeveloped areas	−0.218 **	−0.072 **	−0.126 **	−0.299 **	−0.312 **	−0.072 **	−0.028 **	−0.266 **

Note: ** p < 0.01.

shows that all of the indicators of the HRQoL of the elderly in economically underdeveloped areas deteriorated significantly over the years. On the other hand, only four of the HRQoL indicators—bodily pain, vitality, role-emotional, and mental health of the elderly in developed areas—degenerated over time; the other four indicators of HRQoL improved over time.

Similarly, in order to describe, in detail, the changes in the indicators of HRQoL in different regions over the years, we calculated the effect size for the two regions.

Table 8. Changes in the indicators of the HRQoL of the elderly in different regions over the years.

Index	Economically Developed Areas				Economically Underdeveloped Areas
	ΔM	SD	d	r2	ΔM	SD	d	r2
PF	24.03	19.21	1.25	0.28	−9.36	19.75	−0.47	0.05
RP	9.27	27.14	0.34	0.03	−3.55	31.69	−0.11	0.003
BP	−7.94	17.24	−0.46	0.05	−7.2	17.61	−0.41	0.04
GH	2.32	16.11	0.14	0.005	−11.54	15.94	−0.72	0.11
VT	−4.27	13.53	−0.32	0.02	−12.04	14.84	−0.81	0.14
SF	2.82	19.26	0.15	0.01	−3.44	19.04	−0.18	0.008
RE	−4.07	26.23	−0.16	0.01	−1.3	33.11	−0.04	0.0004
MH	−2.79	13.09	−0.21	0.01	−10.66	14.35	−0.74	0.12

shows that the physical functioning of the elderly in economically developed areas has changed greatly (the change range in effect size is large). The effect sizes of the changes in role-physical, bodily pain, vitality, and mental health were all small. The effect sizes of the changes in the rest of the indicators of HRQoL were close to small. In economically underdeveloped areas, only vitality recorded an obvious change (the effect size of the decline was large). The size of the effect of the decline in general and mental health was medium. Physical functioning and bodily pain recorded small effect sizes. The rest of the indicators failed to attain a small effect size (please see Table 8).

4. Discussion

4.1. The HRQoL of the Elderly in China Has Risen in Some, but Fallen in Other, Indicators over 20 Years

In this study, we cross-temporally meta-analyzed 45 research reports on measuring the HRQoL of the elderly (36,352 subjects in total) in China with a SF-36 scale from 2002 to 2020. It was found that bodily pain, general health, vitality, and mental health decreased over the years, which may be related to the demographic and epidemiological shifts where the “inverted pyramid” family structure [5] and the surge in non-communicable diseases [4] degrade the HRQoL of the elderly. Urbanization has produced a large-scale rural–urban movement in China (a large number of young people move away from their homes to work in other places). This trend has also resulted in an increasing number of elderly people being left behind in rural areas. The rising number of “empty nesters” [4] living alone [38] further compromises the exercise of “filial piety” and, thus, the traditional Chinese intergenerational support system [39]. The absence of children’s company may deplete the physical and mental health [40,41] and, therefore, the vitality and general health of the elderly.

However, with the advancement in information and communication technologies (ICTs) [42], the “empty nesters” or the elderly people living alone can stay connected with their children and have relatively stable socio-emotional support. The use of smartphones by older people has a significant positive impact on subjective health, especially on mental health [43]. Leveraging the strength of ICT as an enabler of family support and real-time connectivity may promote the social functioning and the emotional wellness of the elderly. Empirical studies have found that the use of mobile Internet applications (Apps), including WeChat, WeChat Moments, and mobile payment, can effectively promote the physical and mental health of the elderly. Further evidence shows that the elderly with lower education levels even benefit more from the use of these apps [43].

Despite the inspiring results of the support of ICT for the elderly’s everyday lives, it is important to consider that the elderly may have some difficulties in using smart devices, such as smartphones, especially alone. Additionally, regional differences in ICT infrastructure development may also limit how beneficial ICT is to the elderly. Therefore, further study is necessary to first identify the challenges that the elderly living alone face in using ICT products, such as mobile phones and apps, to inform ways to address them. Second, research analysis of information technology infrastructure development, such as the Internet, especially in remote areas, is urgently needed. The findings of such studies may accelerate the transformation of the benefits of smartphones for the elderly by improving infrastructure in underserved areas to encourage ICT usage, thereby enhancing the elderly’s social functioning and emotional roles over time.

4.2. The Increasing Dependency Ratio of the Elderly Weakens Their HRQoL

As a social indicator, the dependency ratio of the elderly significantly correlates with the HRQoL of the elderly negatively. That is, the rising dependency ratio of the elderly in China implies a decline in the elderly’s HRQoL. This indicates that the social and family pension burden is heavier. Specifically, the implementation of the family planning policy (only-child policy) in the late 1970s has generated an aging population far larger than the population they depend on [44]. Many families have a situation where one young family needs to raise children whilst supporting four elderly people. This may lead to two potential outcomes—first, the family’s support for the elderly may further decline and, second, institutional care may be chosen for the elderly. Either way, the HRQoL of the elderly, as evidenced in this meta-analysis, would be worse off.

To address this structural problem, the high dependency ratio of the elderly must be reduced to improve the HRQoL of the elderly. The adjustment in family planning in China is a step in the right direction. That is, the shift in the Government of China’s advocacy for “fewer and better” births to the current relaxation of childbirth, including the introduction of “two children” for couples who are the only children of their parents, to “two children” even if only one of the couple is the only child of his/her parents, and then to the universal “two-child policy”, regardless of whether a couple comprises only children or not, right to the current universal “three-child birth policy”, may remedy the rising dependency ratio. In this way, the support of the elderly can be shared by the children, and the life pressure on the children can be alleviated. This will be helpful in improving not only the HRQoL of the elderly but also that of the people they depend on. Future research that analyses whether the current birth policy is achieving the intended results is needed to forecast the dependency ratio for the foreseeable future and mitigation strategies.

4.3. The HRQoL of the Elderly under Home Care Is Higher Than That of Those under Institutional Care

Our findings indicate that the HRQoL of the elderly under home care mainly shows an improvement over time. This may be mainly because the elderly live with their children and/or spouses. The family economic conditions in China are better now than before. Therefore, families can now afford to nourish the elderly in terms of their bodies and make available relevant entertainment technologies for their emotional and mental well-being, improving their HRQoL [45]. However, due to the “over-care” of the family, the subjective feelings of the elderly may be ignored and the energy of the elderly will be reduced, making them prone to fatigue [46]. This is mainly reflected in the fact that the three factors of physical pain, vitality, and mental health show a downward trend over time. Family and environment play an important role in the care of the elderly in China [47]. Since the development of institutional care for the elderly, the facilities have not been adequate, the relevant systems have been lacking, and the institutions have only been able to provide basic living services without family care. These basic services are too poor in quality and quantity to be able to cater to the mental and emotional needs of the elderly [3,4]. Therefore, we notice that all of the indicators of the HRQoL of the elderly who are under institutional care have greatly declined over time. These results are consistent with prior research that the aged who live with their family are less depressed than those who do not [38].

For the home-based care model, children should pay more attention to their physical pain, vitality, and mental health, such as by providing better medical services, improving the energy of the elderly, and caring for the spiritual needs of the elderly, so as to improve the quality of life of the elderly. For the institutional pension model, the government actively supports and supervises, promotes the elderly care institutions to continuously strengthen hardware and software investment, and improves the level of elderly care services. Elderly care institutions themselves should take the initiative to improve their hardware and software levels to promote the high-quality development of elderly care services. It is also necessary to build a comprehensive talent training and service quality evaluation system for elderly care institutions. Considering the differences in the elderly HRQoL between the elderly under institutional care and those in the home-based care model, future research that examines the details of how the elderly live under these two models may reveal insights that may be beneficial for improving HRQoL of the elderly.

4.4. The Elderly’s HRQoL Indicators Have All Declined in Underdeveloped Areas but Show Mixed Results in Developed Ones

The results of our CTMA also support the literature evidence that the HRQoL of the elderly varies across regions—developed and underdeveloped [23,48]. The findings show that physical functioning, role-physical, overall health, and social functioning in developed areas are on an upward trend. With continuous investments, the medical conditions in economically developed areas are getting better and better; however, the medical resources in economically underdeveloped areas are relatively scarce [4]. The elderly in economically developed areas have better medical treatments, relatively complete service facilities, and social networks [15,23]. The elderly in economically developed areas typically have the capacity to meet their material needs; so, they pay more attention to their emotional and mental health needs [15]. However, all factors in underdeveloped areas show a downward trend over time. This may be because the rural family structure is undergoing structural changes with the rural–urban migration [18,49].

Due to the restrictions of the household registration system and the high cost of health and education services in cities, most migrant workers leave their children in their hometowns to be taken care of by their parents. However, when the left-behind elderly live with their grandchildren, they not only have to farm but also undertake the tasks of care and education for their grandchildren, which increases the physical and mental burden of the elderly. Moreover, most of the elderly people in economically underdeveloped areas are generally not well educated and, therefore, lack awareness of modern educational methods. They often feel inadequate in educating their grandchildren and feel worried that they may be blamed for their grandchildren’s failure to make it in life. In addition, because children are no longer around for a long time, the elderly often lack the tender care of their children. Lack of family comfort leads to a strong sense of loneliness [50], which affects the HRQoL of the elderly. Therefore, the rural areas in economically underdeveloped areas mainly pay attention to survival-oriented old-age resources while the elderly in rural areas in economically developed areas mainly pay attention to development-oriented resources for the age. Future research that explores the intricate details of the specific activities, events, resources, and access that condition differences in elderly HRQoL in developed and underdeveloped areas would be a useful starting point in helping those in underdeveloped areas.

5. Limitations

In spite of the theoretical and practical implications of this study, a number of limitations are worth mentioning. First, our results on the health-related quality of life (HRQoL) of the elderly in China over a 20-year period were obtained through a cross-temporal meta-analysis of 45 studies that used the SF-36 scale. This scale assesses individual physical and mental factors but does not cover individual cognition, consciousness, etc. The dimensional limitation of this scale limits the scope of generalizability of the research results. Therefore, there is a need to examine other scales that measure other dimensions of elderly quality of life, including the PERMA or PANAS models. Second, our analysis did not account for key demographic characteristics of the elderly that may be informative, such as the economic status of the elderly, whether they are single, and whether they are migrant elderly following their children. This limits the contextual relevance of our analytical results beyond the general trends. Therefore, it is suggested that future studies should pay attention to how the quality of life of the elderly differs across socio-demographic characteristics. Such a deep analysis may reveal ways interventions may be most effective and useful.

6. Conclusions

The HRQoL of the elderly in China is an issue of national concern and is at the heart of nation building. Research reports cast an incomplete picture of the effects of the efforts of governmental and non-governmental agencies on the HRQoL of the elderly [15]. To resolve the inconsistency, we cross-temporally meta-analyzed 45 research reports measuring the HRQoL of the Chinese elderly (36,352 subjects in total) with a SF-36 scale from 2000 to 2020 to answer three main questions. First, is the HRQoL of the elderly in China rising or falling over the years (i.e., from 2000 to 2020)? Our results showed that whilst half of the indicators of HRQoL—bodily pain, general health, vitality, and mental health—are deteriorating, the other three—physical and emotional roles and social functioning—show improvement over time. Second, to what extent does this change (rise or fall) differ across the different elderly care models adopted and the regions of domicile? We found the HRQoL of the elderly under home care to be better than that of those under institutional care. Regionally, the elderly in economically developed areas were recorded as having a higher HRQoL than those in economically underdeveloped areas. Lastly, how does the rising dependency ratio impact the HRQoL of the elderly over the years? Our findings revealed that the rising dependency ratio has a deteriorating effect on the HRQoL of the elderly over time. These findings suggest that there is an urgent need for policies and programs that improve the living resources of the elderly, particularly those in economically underdeveloped areas and/or institutional care. Also, policies that encourage childbirth may decrease the dependency ratio of the elderly, thereby improving family support systems under the traditional Chinese intergenerational support contract system known as “filial piety”.