Community-embedded eldercare services and wellbeing of urban older adults aged 60 to 69 in Shanghai

Abstract

Community-embedded eldercare is a new approach to eldercare that allows older adults to remain embedded in their communities while giving them access to professional care, thus seemingly combining the best of home care and institutional care. The article reports and analyzes the findings from an interview- and questionnaire-based study of urban older adults aged 60 to 69 receiving community-embedded eldercare in Shanghai, focusing on its impact on their wellbeing and potentially mediating factors. The findings are analyzed by means of cognitive appraisal theory and structural equation models. They confirm that both community-embedded eldercare as such and perceptions of it influence older adults’ wellbeing positively, highlighting the role of intergenerational support, community support, community interaction, and economic support, thus also corroborating theoretical assumptions that older adults tend to give more importance to intrinsic rather than instrumental values and that they come to focus more strongly on the emotional aspects of their life.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-14914-z.

Introduction

China is undergoing rapid population aging. According to the seventh national census individuals aged 60 and older account for 18.7% of the population, which surpasses the widely recognized threshold of 10% for an aging society¹. It is estimated that this demographic will constitute roughly 35% by 2050². This burgeoning older population presents unparalleled challenges, necessitating expanded health and long-term care (LTC) systems, which increasingly place pressure on Chinese families.

Traditional informal and formal approaches to elder care in China have been predominantly confined to home-based (Or family care, deeply rooted in the Confucian value of filial piety, has traditionally been the cornerstone of security for old adults in China^3–5) and institutional-based care (This approach typically reserves institutions care for older adults without families, and/or with developmentally disabled issues⁶.), each presenting distinct challenges. Previous studies^7–11 indicate that while home-based care offers a familiar and comforting environment, it may lack professional medical services, institutional care, on the other hand, although professional, can be expensive and disconnect older adults from their families, with potential mismatches between supply and demand. To address these issues, a novel community-based care (The Chinese government has developed various strategies to utilize governmental resources to support aging within community settings^12–15.) model was introduced^16–18, theoretically enabling older adults to access professional care without leaving their communities (These efforts align with the current aging policy of ‘aging in place’, which aims to maintain older adults with low dependency in their own homes and communities, turning to institutional care only when absolutely necessary^18,19.). In Western contexts, this model has shown benefits for the physical and mental health of older adults^20,21, and helps maintain their quality of life in familiar surroundings²². However, the application of this model in China faces challenges, partly due to differences in service provision, services are less comprehensive and more dispersed, often reliant on government resources²³. Recent initiatives are aiming to integrate non-governmental sectors, fostering a community-embedded care (The community-embedded care model represents an innovative evolution of traditional elderly care practices. It integrates professional eldercare workers and community staff as the principal caregivers, positioning them at the forefront of service delivery. In this framework, the government and families are identified as the primary bodies responsible for oversight and support. This model is predominantly community-centric, with home-based care serving as secondary, supportive function. Financially, it is predominantly sustained by government funding, complemented by contributions from families. A distinguishing feature of this model is its emphasis on providing services that are both more personalized and varied, setting it apart from conventional eldercare approaches^4,24 approach influenced by government policies. While this approach is gaining traction, especially in megacities such as Shanghai, (Shanghai, stands out not merely as a significant economic hub, but also China’s ‘oldest city’ in terms of its demographic composition^25–27. It is viewed as a frontrunner in adopting community-embedded elder care services in China.) empirical research on its effectiveness and impact on older adults’ wellbeing remains scarce.

This study addresses that gap by investigating how community-embedded elder care services influence the wellbeing of older adults in Shanghai, one of the most rapidly aging cities in China^25,27. Drawing on Cognitive Appraisal Theory^28,29, the study proposes a Perception-Cognition-Evaluation framework to explore how older adults assess their care environment and how these assessments relate to their wellbeing. (Given the potential for older adults to face more significant challenges in achieving a ‘good life’ compared to younger individuals, it is crucial to prioritize their wellbeing and quality of life³⁰. While older adults can be also resourceful and resilient³¹, Mau et al.³², they tend to be a particularly vulnerable group, facing special physical and not least mental health risks^33–35.

Current research predominantly examines how older adults’ wellbeing (OWB) is affected by various factors such as physical and mental health, socio-economic status^36–38, physical activity^39,40, social networks, family relations, living arrangements, and social participation^41–45. These studies also explore the needs of older adults in traditional home-based, institutional, and community-based settings^46–49. However, there is a notable gap in research regarding their wellbeing within community-embedded care models. Exploring the mechanism and influencing factors of older adults within the context of community-embedded model is not only of academic interest but also essential for practical strategies to improve their life quality.

To fill this gap, this study employs a mixed-methods approach. Primary data were collected from selected communities in Shanghai through semi-structured interviews and survey questionnaires. Structural equation modeling is then applied to examine the relationship between community-embedded elder care services and older adults’ wellbeing. The study contributes new empirical insights to both academic discourse and policy development concerning aging and community care in urban China.

Theories and hypotheses

Older adults’ wellbeing

We understand wellbeing in the most general sense as state or condition which is intrinsically good for a person⁵⁰. More specifically, we employ a notion of subjective wellbeing, taking this to comprise a favorable balance of positive over negative emotional states as well as a positive assessment of one’s life and situation (i.e. “cognitive wellbeing”). This is roughly equivalent to the construct of subjective wellbeing (“SWB”) championed by Ed Diener and widely used in contemporary wellbeing research^51,52). Working with a notion of subjective wellbeing is particularly relevant in the present case, as the question is very much how older adults respond to and “take in” a seemingly well-designed and supportive environment, viz. how they translate objective features of the environment into subjective qualities. There is reason to assume that cultural background⁵³, expectations, norms and even personal characteristics impact on the way the environment is experienced. The focus on emotions is particularly relevant to the study of older adults’ wellbeing, as older people are known both to be generally good at emotional regulation but also be able to hide emotional distress even from themselves, as some older adults can at the time show symptoms of depression and tend to judge their life positively³⁴. This is also a reason to avoid relying exclusively on self-reporting, in contrast to much existing wellbeing research.

Cognitive appraisal theory

Cognitive appraisal theory is a general type of psychological theory that describes the subjective interpretation of stimuli in the environment, notably the way individuals respond to “stressors”, that is, factors that seem demanding, challenging or threatening to them^28,29. This theory fits well with the subjective wellbeing approach and specific theories of older adults’ wellbeing³⁴, which it supports and complements. It assumes that personality and values, but also specific coping skills or perceptual dispositions, mediate factors in the environment. The coping mechanism it posits—for example, focusing on positive aspects of a stimulus⁵⁴, broadening one’s perspective⁵⁵ or reappraising an event⁵⁶—closely match what has been, in wellbeing theory, described as mechanisms for hedonic adaptation in a broad sense⁵⁷, for similar considerations of the reappraisal of past events, see Klausen and Hasandedic-Dapo⁵⁸.

Coping is not just a matter of emotional regulation or selective attention, however. Certain environment facilitates the use of coping strategies among older adults, enhancing their ability to respond adaptively to stressors. Especially various forms of social support can help individuals to focus on the positive aspects of a situation or broadening their perspective⁵⁷. Hence the mediators are not purely intrapsychic, even on a subjective wellbeing and cognitive appraisal approach. Personality traits may be fixed, but their effect can differ depending on the context. Values and perceptual dispositions can likewise be modified, or their adverse consequences mitigated, due to social influences and eldercare interventions. Consequently, variables such as eldercare support and social interaction should not merely be treated as background conditions but as potential mediating factors in how older adults cognitively and emotionally assess their lives.

Building upon these theoretical foundations, this study adopts a Perception-Cognition-Evaluation model to conceptualize how older adults in community-embedded eldercare settings experience and evaluate their wellbeing (see Fig. 1). The model consists of three interconnected dimensions: (1) Perception refers to how older adults subjectively experience the care environment, encompassing their impressions on service quality, emotional atmosphere, and social embeddedness; (2) Cognition reflects the internal processing of these perceptions, including evaluative judgments like life satisfaction, acceptance, or meaning-making; (3) Evaluation involves the outcome-level appraisals or emotional states that emerge from this process, including contentment, happiness, loneliness, or anxiety.

Fig. 1.

Theoretical model.

This tripartite model allows for an integrated analysis of how older adults translate their lived experience into broader evaluations of wellbeing, and how various social and institutional supports influence this trajectory. The following figure illustrates the proposed conceptual framework.

Community-embedded eldercare and older adults’ wellbeing

Community-embedded eldercare refers to a model where eldercare services are deeply integrated into the community’s social and physical infrastructure, allowing various stakeholders (local government, local businesses, families and non-profits) to collaborate and older adults to receive care services in familiar settings, thereby enhancing their quality of life^23,24. Studies have indicated that community-based eldercare is closely associated with older adults’ wellbeing. By providing a familiar environment and support network, this care model can reduce loneliness and anxiety among them, while enhancing their self-esteem and life satisfaction^59,60. Additionally, community-embedded care can foster older adults’ social engagement, increasing their interaction and connection with others, thereby improving their happiness⁶¹.

Based on Cognitive Appraisal Theory, older adults’ wellbeing in a community-embedded elder care environment can be analyzed with a Perception-Cognition-Evaluation model (see Fig. 1). In this study, perception is understood as older adults’ subjective experiences of being in a community-embedded elder care environment, including their impressions of service quality, the availability and reliability of support systems, and their emotional responses to daily life. These perceptions are assumed to influence cognitive appraisals, such as judgments about life satisfaction, and to directly shape enduring emotional states, both positive and negative, like contentment, joy, loneliness, or depression.

This definition lays the foundation for classifying “eldercare perception” and “eldercare support” as part of the “Perception” dimension. Eldercare perception refers to the subjective impressions and emotional responses of older people to the care services, emphasizing the direct interactions within these environments and direct experience with the quality-of-care services. On the other hand, Eldercare support, which addresses older adults’ direct experience with support systems, is an umbrella term that covers a range of activities aimed at helping older adults, including family support and social support. Peng et al.²³ categorizes intergenerational family support into financial contributions, daily caregiving and emotional interaction, noting that these support systems significantly bolster older adults’ wellbeing by fulfilling crucial emotional and practical needs. With regard to social support, it was segmented by scholars into four subtypes: emotional, informational, instrumental and companionship, each contributing positively to older people’s wellbeing^62,63. It was similarly supported by Chen⁶⁴ that basic pensions generally exert a positive influence on older adults’ wellbeing, with notable disparities between rural and urban cohorts though.

Based on a synthesis of the existing findings and assumptions, the following research hypotheses are proposed:

H1a

Eldercare perception of community-embedded elder care services positively promotes older adults’ wellbeing.

H1b

Eldercare support within community-embedded elder care services positively promotes older adults’ wellbeing.

Community-embedded care and eldercare satisfaction

Life satisfaction is understood within the cognitive domain as an individual’s assessment of the discrepancy between the “ideal” and “actual” conditions of various life aspects⁶⁵. This study primarily examines the perceptions of life satisfaction among older adults in community-embedded eldercare facilities, subsequently categorizing this dimension as eldercare satisfaction.

Previous studies identify multiple factors influencing life satisfaction among older adults, including demographic characteristics, health status, economic conditions, formal social support systems and characteristics of offspring^11,66–69. Research findings in this field exhibit diversity among scholars; however, there is consensus that certain elements positively impact eldercare satisfaction. Active social engagement^45,70, comprehensive healthcare services^12,71, robust social support^72,73, and good physical and mental health have been consistently found to enhance satisfaction among older adults^74,75. These elements frequently intersect with factors contributing to the subjective wellbeing of this demographic group, underscoring the complexity and multidimensionality of factors influencing their satisfaction.

In this paper, Eldercare satisfaction specifically refers to older adults’ overall cognition of the community-embedded care approach during their immersive experience. This includes external dimensions such as the natural environment, the cultural environment of the community, and service provision, as well as internal dimensions such as the older adults’ perceived sense of autonomy, emotional comfort, and capacity to engage meaningfully with the community. Although individual conditions, such as health status or psychological resilience, may influence how older adults interpret their care experience, these are not considered intrinsic components of eldercare satisfaction in our model. Rather, they are treated as antecedent or moderating variables, consistent with prior studies on cognitive appraisal in eldercare contexts^45,75.

The following hypotheses are proposed,

H2a

Eldercare perception of community-embedded elder care services can improve eldercare satisfaction.

H2b

Eldercare support within community-embedded elder care services can improve eldercare satisfaction.

Eldercare satisfaction and older adults’ wellbeing

This research is centered on the community-embedded eldercare model, where eldercare satisfaction specifically denotes older adults’ comprehensive cognitive perception of eldercare facilities, evaluated through immersive experiences. This satisfaction encompasses assessments of the natural environment, the sociocultural environment of the community, and individual personal conditions. Happiness, positioned within the emotional domain, is conceptualized as an individual’s emotional assessment of life, encompassing feelings, experiences, and evaluations⁷⁶.

Subjective wellbeing comprises three components: positive affect, negative affect, and life satisfaction. Research on subjective wellbeing has bifurcated into two predominant orientations: quality of life and psychological health, with a growing tendency towards integration⁷⁷. Typically, subjective wellbeing is characterized as an individual’s cognitive evaluation of life satisfaction. As individuals experience heightened satisfaction, their evaluative judgments of their lives improve, thereby enhancing overall happiness.

The following hypothesis is proposed:

H3

Eldercare satisfaction in community-embedded elder care services positively influences older adults’ wellbeing.

Eldercare satisfaction serves as a mediator between community-embedded eldercare and older adults’ wellbeing (see Fig. 2 research model)

Fig. 2.

Community-embedded eldercare service and older adults’ wellbeing research model.

According to the Perception-Cognition-Evaluation theoretical model, individuals form positive perceptions in response to external stimuli, which then leads to positive cognitions and subsequent evaluations. In quantitative assessments, this model incorporates objective indicators reflecting conditions of eldercare, referred to as the “inputs” of eldercare quality. It also encompasses subjective indicators that capture the levels of satisfaction, defined as the “outputs” of life quality. These assessments are ultimately combined by the model to effectively bridge objective conditions and subjective wellbeing in eldercare settings.

The following hypotheses are proposed:

H4a

Eldercare satisfaction mediates the relationship between eldercare perception and older adults’ wellbeing.

H4b

Eldercare satisfaction mediates the relationship between eldercare support and older adults’ wellbeing.

Methods

This study utilized a mixed-methods approach, integrating both qualitative and quantitative techniques. For qualitative data collection, semi-structured interviews and open-ended questions were employed, while a questionnaire was used to gather supplementary quantitative data. Although the quantitative analysis might be considered sufficient to test our hypotheses and overall research model, the qualitative part was added to gain knowledge about how the mediation was done (or not done), and how the objective conditions impacted on subjective wellbeing, that is, how far they were actually experienced as relevant to wellbeing (compare 2.1 above on the translation of objective factors into subjective qualities).

Interview guide

The qualitative research primarily relied on in-depth interviews conducted with retired older adults having lived experiences with community-embedded care facilities in Shanghai. A total of nine older individuals participated in these comprehensive interviews. All participants volunteered for the study, and the interview content, designed with reference to relevant literature, included the following seven questions in addition to general demographic characteristics: “Do you think the current method of elder care suitable for your needs?,” “What changes have you observed in your daily life or psychological state since participating in community-embedded elder care?,” “In your opinion, what aspects of the community-embedded elder care facilities are well-managed?,” “What aspects of the community-embedded elder care facilities need improvement?,” “Do you feel supported by your family in your participation in community-embedded elder care?,” “What aspects of your life bring you the most happiness?,” “Have you experienced any recent worries or concerns?” Because the qualitative part was not intended as a contribution to our hypothesis testing, the questions were not directly informed by Cognitive Appraisal Theory, but rather by general subjective wellbeing theory and, more specifically, by our interest in understanding how objective factors in the environment are experienced as impacting on the older adults’ wellbeing. Also because the aim was merely to add an understanding of how the mediation might play out as seen from a more subjective perspective, the nine interviews—which were quite comprehensive—were considered more than sufficient to achieve saturation.

The interviews, conducted in Chinese and later translated into English for analysis, each lasted between 30 and 60 min. To analyze the data, a descriptive, inductive, and conventional content-analysis approach was used^78,79. This method is notable for its effort to draw conclusions directly from the text, minimizing the use of pre-existing categories and hypotheses and allowing categories to naturally emerge from the data. The interview content was then aligned with the theoretical framework of the research and used to formulate the questionnaire items.

Questionnaire

In this study, we used a custom-designed questionnaire to gather primary data on older adults’ wellbeing, who have lived experiences with community-embedded eldercare facilities. Based on the interview results and a comprehensive review of existing literature, the questionnaire for this study is divided into nine dimensions: Planning Perception (PP), Environmental Perception (EP), Service Perception (SP), Intergenerational Support (INGS), Community Support (CS), Interaction Support (INAS), Economic Support (ES), Eldercare Satisfaction (SA), and older adults’ wellbeing (OWB). The three items for older adults’ wellbeing were referenced from the study conducted by Sapin et al.⁸⁰: 1. Frequency of feeling unhappy and depressed in the past four weeks, 2. Frequency of feeling overwhelmed by problems in the past four weeks, 3. Overall satisfaction with current life. Each aspect scored on a numerical 7-point Likert scale⁸¹, ranging from “strongly disagree” to “strongly agree”, with a higher score reflecting a higher level of wellbeing.

The dimensions of planning perception, environmental perception, and service perception originate from older adults’ intrinsic subjective consciousness, offering a personal evaluation of the software and hardware of current community-embedded eldercare institutions. These three dimensions are collectively referred to as Eldercare Perception. The following items were designed based on interview content and previous research findings^24,49. For planning perception, they were: 1. There is convenient parking at the community-embedded eldercare institution, 2. Good adaptation and accessibility modifications at the community-embedded eldercare institution, 3. Comprehensive supporting facilities at the community-embedded elder care institution, 4. Ample activity space and fitness equipment at the community-embedded eldercare institution. For environmental perception, they were: 1. Fresh air at the community-embedded eldercare institution, 2. Good greenery at the community-embedded eldercare institution, 3. Clean surrounding areas of the community-embedded eldercare institution. For service perception, they were:1. Well-managed community-embedded eldercare institution, 2. Good service work by the residents’ committee at the community-embedded eldercare institution, 3. Excellent medical services at the community-embedded eldercare institution, 4. High-quality service from the staff at the community-embedded eldercare institution.

Intergenerational Support, Community Support, Interaction Support, and Economic Support focus on the objective evaluation of external conditions affecting older adults’ participation in community-embedded eldercare. These four dimensions can be collectively termed Eldercare Support. The following items were developed based on interview content and previous research findings^82–85. For intergenerational support, they were: 1. Close relationship with children, 2. Children encouraging participation in community-embedded eldercare institutions, 3. Financial support from children for participation in community-embedded eldercare institutions, 4. Participation in community-embedded eldercare institutions making children’s lives easier. For community support, they were: 1. Frequent volunteer visits to the community-embedded eldercare institution, 2. Frequent visits from the residents’ committee staff to explain eldercare policies, 3. Certain financial subsidies provided by the community, 4. Frequent sight of public service posters at the community-embedded eldercare institution. For interaction support, they were: 1. Willingness to participate in activities at the community-embedded eldercare institution, 2. Making new friends at the community-embedded eldercare institution, 3. Increased visits to neighbors in the past two weeks, 4. Increased visits from neighbors in the past two weeks. For economic support, they were: 1. Acceptable costs of the community-embedded eldercare institution, 2. Different fee standards for different service needs, 3. Humane fee structure of the community-embedded eldercare institution.

Questions about eldercare satisfaction centered around: 1. The community-embedded eldercare institution has many advantages, 2. Agreement with the community-embedded eldercare model, 3. The community-embedded eldercare model is more efficient than traditional eldercare. These items were designed based on interview content and relevant literature^86,87.

Additionally, the study also collected general demographic characteristics of older adults, including gender, age, education, marital status, living conditions, number of children, and monthly income. On average, each participant took 30 min to complete the questionnaire. To guarantee the reliability and objectivity of the responses, participants were assured of their anonymity. They were informed that there were no right or wrong answers, and were encouraged to answer truthfully based on their own experiences and situations.

Data collection and analysis

Before initiating the formal survey, a pilot study was conducted to ensure the reliability and validity of the questionnaire. The pilot study was carried out in June 2024 in Xuhui District. A total of 132 questionnaires were distributed, and 98 were collected. The pilot study data revealed that Cronbach’s alpha values for each dimension were all above 0.785, indicating good reliability. Additionally, factor analysis showed that the factor loadings for each item were above 0.483, demonstrating good construct validity. Therefore, the questionnaire used in this study possesses satisfactory reliability and validity, making it suitable for formal research.

This study focused on older adults aged 60–69, aiming to explore the subjective wellbeing of younger older adults who have experiences with community-embedded eldercare and its influencing factors. The formal study selected 6 districts across Shanghai. Data collection was conducted between July and August 2024, employing a mixed-method approach that combined on-site paper-based questionnaires with online surveys. The paper-based questionnaires were collected by students at each community-embedded eldercare institution or the relevant community, with each elderly participant completing the questionnaire through one-on-one interviews. Respondents who refused to participate, had language barriers, or could not fully understand the questionnaire were excluded. Online surveys utilized for data collection was SurveyStar. Out of the initially distributed 543 questionnaires, 495 were completed with logical consistency, resulting in a response rate of 91.16% (495/543). The regional distribution of respondents was as follows: Pudong (82 participants, 15.10%), Minhang (78 participants, 14.36%), Xuhui (94 participants, 17.31%), Baoshan (83 participants, 15.28%), Qingpu (72 participants, 13.25%), and Songjiang (86 participants, 15.83%). This aligns with the recommended general questionnaire sample size⁸⁸, which suggests that at least 10 responses per estimated parameter are sufficient for analysis⁸⁹.

For analyzing the survey data, the structural equation model (SEM) was employed to investigate the relationships between community-embedded eldercare services, eldercare satisfaction, and older adults’ wellbeing. Given the conceptual structure of the model, including several second-order latent constructs, a two-step modeling strategy was adopted, following Anderson and Gerbing⁹⁰. SPSS 25.0 and MPLUS8.0 were used as statistical software. In the first step, we conducted Confirmatory Factor Analysis (CFA) to validate the measurement model, ensuring that each latent construct was reliably measured by its observed indicators. We evaluated convergent and discriminant validity using factor loadings, Average Variance Extracted (AVE), and Composite Reliability (CR). Model fit was assessed using multiple fit indices, including the Chi-square/df ratio (χ²/df), Comparative Fit Index (CFI), Tucker-Lewis (TLI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR). These indices will be reported as obtained directly from the initial model specificatoin, without post hoc modifications such as correlating error terms or removing items, in order to maintain conceptual integrity and analytic transparency.

In the second step, we tested the structural model to assess the hypothesized relationships among latent constructs. The Maximum Likelihood Estimation (MLE) method was employed, given its robustness and wide applicability in SEM, and its suitability for data that approximate multivariate normality. Covariates such as age, gender, educational attainment, marital status, number of children, living arrangement and monthly income were included as exogenous control variables. Paths were specified from these covariates to the primary outcome variable (older adults’ wellbeing) and, where appropriate, to eldercare satisfaction, in order to control for potential confounding effects. Prior to SEM, all data were screened for missing values, outliers, and normality. The final sample consisted of 495 valid responses, and all variables met acceptable thresholds for skewness and kurtosis, supporting the use of MLE.

Findings

Descriptive analysis

A total of 495 valid questionnaires were collected for analysis. The gender distribution revealed a higher proportion of females (61%) compared to males (39%) (see Table 1). In terms of educational attainment, a significant proportion of respondents had an education level of elementary school or below (44%) and middle school (39%). Marital status analysis indicated that the proportion of married elderly individuals (52.9%) was comparable to those with other marital statuses (47.1%). Regarding the number of children, a notable proportion of the elderly had either one child (29.7%) or two children (34.9%). Analysis of living arrangements showed that a larger proportion of elderly individuals lived independently from their children or grandchildren (57.8%) compared to those who cohabited with them (47.2%). Monthly income analysis revealed that the highest proportion of respondents had a monthly income of 3001–5000 yuan (30.7%), followed by those with a monthly income of 1001–3000 yuan (29.1%).

Table 1.

Socio-demographic information of the participants in the survey.

Type	Variable	Frequency	Percentage
Gender	Male	193	39.0
	Female	302	61.0
Age	60	42	8.5
	61	42	8.5
	62	38	7.7
	63	60	12.1
	64	57	11.5
	65	61	12.3
	66	56	11.3
	67	49	9.9
	68	36	7.3
	69	54	10.9
Education	Primary/lower	218	44.0
	Junior high school	193	39.0
	Senior high school/vocational/technical	53	10.7
	College/academic higher education	31	6.3
Marital status	Married	262	52.9
	Other	233	47.1
N of children	0	45	9.1
	1	147	29.7
	2	173	34.9
	3	76	15.4
	> 4	54	10.9
Living status	Living with children	209	42.2
	Living alone	286	57.8
Income	< 1000	110	22.2
	1001–3000	144	29.1
	3001–5000	152	30.7
	5001–10,000	89	18.0

Common method bias

If common method bias exists in the data, it can result in the emergence of misleading or artificial relationships between constructs. The common method bias was assessed using HARMAN’s single-factor test in this study. The results indicate that the variance explained by the first unrotated factor was 36.897%, which is below the critical threshold of 50%^91,92. Additionally, the total variance explained was 81.793%. These findings suggest that common method bias is not a significant concern in this study.

Reliability analysis

The reliability of this study was assessed using the Cronbach’s Alpha coefficient. Rule of thumb suggests that a Cronbach’s Alpha value above 0.7 for multi-item constructs indicates high internal consistency, suggesting that the scale is reliable and suitable for further analysis^93,94. If the value falls below 0.7, the questionnaire requires revision⁹⁴. In this study, the reliability coefficients for each dimension are provided in Table 2. It reveals that Cronbach’s Alpha coefficients for all variables are greater than the standard threshold of 0.7, indicating that the variables exhibit good internal consistency reliability.

Table 2.

Reliability analysis.

Variable	Cronbach’s alpha	N of items
PP	0.940	4
EP	0.882	3
SP	0.907	4
INGS	0.928	4
CS	0.892	4
INAS	0.935	4
ES	0.867	3
SA	0.897	3
OWB	0.937	4

PP, planning perception; EP, environmental perception; SP, service perception; INGS, intergenerational support; CS, community support; INAS, interaction support; ES, economic support; SA, eldercare satisfaction; OWB, older adults’ wellbeing.

Exploratory factor analysis

The construct validity of the questionnaire was analyzed using exploratory factor analysis which was conducted using SPSS 25.0. The KMO test and Bartlett’s test of sphericity were employed to assess the suitability of the scale. As detailed in Table 3, the KMO value is 0.933, exceeding the threshold of 0.7. The Bartlett’s test of sphericity is significant (Sig. < 0.001), indicating that there are common factors among the correlation matrices, confirming that the questionnaire data meet the requirements for factor analysis. Therefore, further analysis was conducted using principal component analysis for factor extraction, with eigenvalues greater than 1 as the criterion for selecting factors. Varimax orthogonal rotation was employed during factor analysis. The results of the analysis are shown in the table below:

Table 3.

KMO and Bartlett’s test.

Kaiser–Meyer–Olkin measure of sampling adequacy		0.933
Bartlett’s test of sphericity	Approx. Chi-square	13,673.970
	df	528
	Sig	0.000

The table (Table 4) indicates that nine factors were extracted from the factor analysis, with a total explained variance of 81.793%, which is well above the 50% threshold. This demonstrates that the nine extracted factors have strong representativeness.

Table 4.

Total variance explained.

Component	Initial Eigenvalues			Extraction sums of squared loadings			Rotation sums of squared loadings
	Total	% of variance	Cumulative %	Total	% of variance	Cumulative %	Total	% of variance	Cumulative %
1	12.176	36.897	36.897	12.176	36.897	36.897	3.427	10.384	10.384
2	4.533	13.737	50.633	4.533	13.737	50.633	3.391	10.275	20.659
3	2.450	7.425	58.058	2.450	7.425	58.058	3.335	10.106	30.765
4	1.605	4.864	62.922	1.605	4.864	62.922	3.301	10.004	40.768
5	1.463	4.433	67.355	1.463	4.433	67.355	3.264	9.891	50.659
6	1.373	4.162	71.517	1.373	4.162	71.517	3.122	9.460	60.119
7	1.244	3.770	75.287	1.244	3.770	75.287	2.457	7.444	67.563
8	1.145	3.470	78.757	1.145	3.470	78.757	2.360	7.151	74.714
9	1.002	3.036	81.793	1.002	3.036	81.793	2.336	7.079	81.793
10	.446	1.352	83.145
11	.410	1.242	84.387
12	.376	1.138	85.525
13	.343	1.039	86.564
14	.329	.996	87.561
15	.313	.947	88.508
16	.308	.932	89.441
17	.297	.901	90.342
18	.284	.860	91.202
19	.269	.816	92.018
20	.255	.773	92.791
21	.248	.750	93.541
22	.238	.721	94.262
23	.228	.691	94.954
24	.215	.651	95.605
25	.201	.608	96.213
26	.195	.592	96.804
27	.181	.549	97.353
28	.171	.517	97.870
29	.166	.502	98.372
30	.162	.491	98.863
31	.139	.420	99.283
32	.132	.401	99.683
33	.104	.317	100.000

The factor loadings are presented in the table above (Table 5). Each measurement item exhibits a factor loading greater than 0.5, while the cross-loadings are all below 0.4. Additionally, each item loads appropriately onto its corresponding factor, demonstrating that the scale possesses good construct validity (Table 6).

Table 5.

Rotated component matrix.

	Component
	1	2	3	4	5	6	7	8	9
OWB4	0.853
OWB1	0.848
OWB3	0.817
OWB2	0.809
PP2		0.857
PP1		0.844
PP3		0.815
PP4		0.791
INGS1			0.835
INGS3			0.824
INGS2			0.814
INGS4			0.797
SP1				0.839
SP4				0.817
SP3				0.800
SP2				0.798
INAS1					0.823
INAS4					0.813
INAS2					0.809
INAS3					0.782
CS4						0.816
CS1						0.795
CS2						0.776
CS3						0.738
SA3							0.832
SA2							0.824
SA1							0.805
ES2								0.830
ES1								0.794
ES3								0.792
EP1									0.811
EP2									0.795
EP3									0.783

Extraction method: principal component analysis.

Rotation method: Varimax with Kaiser Normalization.

1 = Older adults’ wellbeing; 2 = Planning Perception; 3 = Intergenerational Support; 4 = Service Perception; 5 = Interaction Support; 6 = Community Support; 7 = Eldercare Satisfaction; 8 = Economic Support; 9 = Environmental Perception.

Table 6.

Convergent validity.

Variable	Item	Factor loading	CR	AVE
PP	PP1	0.930	0.942	0.802
	PP2	0.932
	PP3	0.885
	PP4	0.831
EP	EP1	0.851	0.883	0.716
	EP2	0.827
	EP3	0.861
SP	SP1	0.911	0.908	0.711
	SP2	0.821
	SP3	0.807
	SP4	0.831
INGS	INGS1	0.922	0.929	0.767
	INGS2	0.863
	INGS3	0.855
	INGS4	0.861
CS	CS1	0.884	0.893	0.676
	CS2	0.783
	CS3	0.779
	CS4	0.839
INAS	INAS1	0.903	0.936	0.785
	INAS2	0.869
	INAS3	0.868
	INAS4	0.903
ES	ES1	0.903	0.868	0.687
	ES2	0.779
	ES3	0.800
SA	SA1	0.901	0.899	0.748
	SA2	0.796
	SA3	0.893
OWB	OWB1	0.930	0.938	0.792
	OWB2	0.831
	OWB3	0.888
	OWB4	0.908

CR, composition reliability; AVE, average variation extraction.

Confirmatory factor analysis

All variables were modeled using MPLUS 8.0, and their validity was evaluated through confirmatory factor analysis (CFA), following the two-step approach before testing the hypotheses⁹⁰. The CFA process is guided by theoretical relationships among both observed and latent variables⁹⁵. The modeling diagram is provided in Table 7.

Table 7.

Discriminant validity test.

	ELCP	ELCS	SA	OWB
ELCP	0.790
ELCS	0.359**	0.762
SA	0.427**	0.503**	0.865
OWB	0.407**	0.428**	0.569**	0.890

The bold diagonal elements represent the square roots of each AVE, while the off-diagonal elements display the construct correlations. **p < 0.01.

Before assessing the goodness-of-fit, we evaluated the measurement model’s convergent and discriminant validity due to the significant role of validity in examining measurement models. Fornell and Larcker⁹⁶ established that three criteria must be satisfied to confirm convergent validity: (a) all factor loadings for the latent variables must be significant and exceed 0.7, (b) the composite reliability (CR) for each construct must be greater than 0.7; and (c) the average variance extracted (AVE) must be above 0.5. Meeting these criteria demonstrates that each variable possesses good convergent validity^95,97,98. The results presented in Table 6 indicate that the standardized factor loadings for each measurement index of planning perception (PP), environmental perception (EP), service perception (SP), intergenerational support (INGS), community support (CS), interaction support (INAS), economic support (ES), eldercare satisfaction (SA) and older adults’ wellbeing (OWB) range from 0.779 to 0.932, the CR ranges from 0.868 to 0.942, and the AVE ranges from 0.676 to 0.802, respectively, all of which exceed the recommended thresholds. These findings confirm that each variable has strong convergent validity and that all hypothesized constructs are supported by reliable indicators.

Discriminant validity refers to the extent to which a construct is distinct from other constructs, indicated by the construct’s shared variance with its indicators being higher than its shared variance with other constructs^97,99. Discriminant validity was assessed by comparing the squared correlations between a given construct and other constructs, along with calculating their average variance. When the square root of the AVE exceeds the off-diagonal values of the relevant variable in the corresponding columns and rows, it indicates that the construct has a stronger association with its indicators than with other constructs^96,100. The correlation matrix test results are presented in Table 7. The bold numbers on the diagonal line represent the square root of the AVE for each construct, while the numbers below the diagonal represent the correlation coefficients between the constructs. In this study, each reflective variable’s value exceeds the off-diagonal squared correlations, demonstrating that the discriminant validity of the research is satisfactory.

Structural equation modeling analysis

Structural equation modeling (SEM) is a widely utilized method that consists of two main components: the measurement model and the structural model. These components are critical in testing the validity of theoretical models involving latent variables^95,101. A complete SEM analysis can be performed only when both components meet the required criteria. In this study, SEM was evaluated and analyzed using MPLUS 8.0, following the assessment of the initial model’s goodness-of-fit, and the fit results are shown in Fig. 3.

Fig. 3.

Hypothesized structural equation model for explaining research use.

To assess the model fit, the structural model was utilized to determine how effectively it represents the data. Several indices are available for this purpose. As noted by Brown¹⁰², the chi-square statistic is highly sensitive to sample size, so it is important to use a range of other fit indices to assess the overall fit of a CFA solution, including the Comparative Fit Index (CFI), the Tucker-Lewis Index (TLI), the root mean square error of approximation (RMSEA) and the square root mean residual (SRMR). Rule of thumb suggests that the chi-square fit statistic/degree of freedom ratio (CMIN/DF) should be below 3, CFI and TLI should exceed 0.9^103–105. Conversely, RMSEA and SRMR should be less than 0.08 to indicate a good fit^103,106,107. As indicated in Table 8, the value of CMIN/DF is 1.236, well within the recommended threshold of 3. CFI and TLI are all surpass the minimum standard of 0.9, with 0.992 and 0.991 respectively. Additionally, RMSEA and SRMR values are 0.022 and 0.037, which are both below 0.08. Therefore, this hypothesized model meets the criteria for a relatively good fit.

Table 8.

Fit indices of measurement and structural model.

CMIN	DF	CMIN/DF	RMSEA	CFI	TLI	SRMR
–	–	< 3	< 0.08	> 0.9	> 0.9	< 0.08
595.536	482	1.236	0.022	0.992	0.991	0.037

CMIN, chi-square fit statistics; DF, degree of freedom; RMSEA, root mean square error of approximation; CFI, comparative fit index; TLI, Tucker Lewis Index; SRMR, square root mean residual.

Model path analysis

Given the strong model fit, the hypothesized relationships among the latent variables were tested using SEM. As demonstrated in Table 9, the results show the standardized path coefficients and their significance levels. The findings indicate that eldercare perception (ELCP) has a significant positive impact on older adults’ wellbeing (OWB) (β = 0.151, p < 0.05), confirming the hypothesis. Similarly, eldercare support (ELCS) also has a significant positive impact on older adults’ wellbeing (OWB) (β = 0.311, p < 0.05), confirming the hypothesis. Eldercare satisfaction (SA) further exhibits a significantly positive impact on older adults’ wellbeing (OWB) (β = 0.389, p < 0.05), confirming the hypothesis. Eldercare support (ELCS) positively influences eldercare satisfaction (SA) (β = 0.339, p < 0.05) as well, confirming the hypothesis. Additionally, eldercare perception (ELCP) significantly contributes to eldercare satisfaction (SA) (β = 0.343, p < 0.05), confirming the hypothesis. Overall, all constructs and hypotheses in this structural model are statistically significant, thereby supporting and validating the hypotheses as reasonably robust.

Table 9.

Model path analysis.

Path	STD. estimate	S.E	Est./S.E	P-value	Result
H1a: ELCP-OWB	0.151	0.050	3.041	0.002	Supported
H1b: ELCS-OWB	0.311	0.048	6.533	0.000	Supported
H3: SA-OWB	0.389	0.046	8.414	0.000	Supported
H2b: ELCS-SA	0.339	0.051	6.647	0.000	Supported
H2a: ELCP-SA	0.343	0.052	6.652	0.000	Supported

ELCP, eldercare perception; OWB, older adults’ wellbeing; ELCS, eldercare support; SA, eldercare satisfaction.

Mediation effect analysis and path comparison

The mediation effect was tested using the BOOTSTRAP method, which generates 1000 resamples of the data within the framework of 95% confidence interval. Based on the findings of MacKinnon et al.¹⁰⁸, the Bias-corrected approach in non-parametric Bootstrap methods is considered the most effective one. The results (Table 10) indicate that the total effect of eldercare perception (ELCP) on older adults’ wellbeing (OWB) is 0.285, with the Bias-corrected 95% confidence interval that excludes zero, indicating the total effect is statistically significant. The indirect effect of ELCP on OWB through eldercare satisfaction (SA) is 0.133, with a confidence interval also excluding zero, thus confirming the significance of the indirect effect. The direct effect of ELCP on OWB is 0.151, and its confidence interval similarly excludes zero, indicating a significant direct effect. These findings suggest that SA partially mediates the relationship between ELCP and OWB, thereby supporting the hypothesis.

Table 10.

Analysis and comparison of mediation effects.

Path	STD. estimate	Bias-corrected 95% CI
		Lower	Upper
Total effects
ELCP-OWB	0.285	0.182	0.383
ELCS-OWB	0.443	0.343	0.535
Indirect effects
ELCP-SA-OWB	0.133	0.082	0.195
ELCS-SA-OWB	0.132	0.083	0.192
Direct effects
ELCP-OWB	0.151	0.049	0.253
ELCS-OWB	0.311	0.211	0.410

ELCP, eldercare perception; OWB, older adults’ wellbeing; ELCS, eldercare support; SA, eldercare satisfaction.

Similarly, the total effect of eldercare support (ELCS) on older adults’ wellbeing (OWB) is 0.443, with a confidence interval that excludes zero, signifying a significant total effect. The indirect effect of ELCS on OWB through SA is 0.132, with a confidence interval that does not include zero, affirming the significance of the indirect effect. The direct effect of ELCS on OWB is 0.311, with its confidence interval excluding zero, indicating a significant direct effect. These results demonstrate that SA partially mediates the relationship between ELCS and OWB, thereby supporting the hypothesis.

Conclusion and discussion

This is a study focused on younger older adults aged 60 to 69, a group primarily in need of basic care, emotional support, and health management, with relatively fewer individuals experiencing severe disabilities. They possess a clear understanding and judgment of their subjective wellbeing within the context of community-embedded eldercare. Additionally, their logical thinking and ability to articulate their thoughts enhance the accuracy and reliability of the survey responses.

Hypothesis 1a, that eldercare perception positively influences the subjective wellbeing of older adults was confirmed. Enhancements in hardware facilities, environmental planning, and service experiences within community-embedded eldercare institutions significantly boost the subjective wellbeing of this cohort, aligning with the findings of other scholars. For instance, Wang and Zhang¹⁰⁹ found that improvements in community infrastructure positively affect residents’ environmental perceptions and their subjective wellbeing. Similarly, Chen et al.⁷, demonstrated that community-embedded eldercare, which integrates family, community, and institutional care, effectively addresses complex issues and meets the service needs of older adults, thereby enhancing their happiness. These results suggest that the widespread promotion and acceptance of community-embedded eldercare depend on robust infrastructure and efficient service management.

Furthermore, the study revealed that increasing intergenerational support, community support, community interaction, and economic support enhances older adults’ wellbeing in community-embedded eldercare institutions. Hypothesis 1b was thus confirmed. Unlike previous analyses, this study explored the impact of external factors from individual, family, and social dimensions. Research by Si et al.¹¹⁰ indicated that higher income levels among older adults improve their happiness, while Wang⁴⁹ suggested that increased community interaction provides emotional comfort, thereby enhancing their wellbeing. Others argued that government guidance, community support, and collaborative efforts from stakeholders and resource-dependent entities are essential to improving the quality of community-embedded eldercare services and earning the favor of older adults. This study also supports these perspectives, highlighting the importance of the joint efforts of individuals, families, and communities in enhancing elder care under the new community-embedded eldercare model.

The study also examined the mediating role of eldercare satisfaction in the relationship between eldercare perception/eldercare support and older adults’ wellbeing. The findings indicate that improvements in eldercare perception and eldercare support positively influence older adults’ satisfaction with community-embedded eldercare institutions. Hypotheses 4a and 4b were thus confirmed. According to the perception-cognition-evaluation model, increased satisfaction further enhances the subjective wellbeing of this cohort, reflecting their true perceptions of community-embedded eldercare institutions, from external observation to subjective cognition and emotional upliftment.

Hypotheses 2a and 2b, that eldercare perception/ eldercare support had a positive impact on eldercare satisfaction were also confirmed. Interestingly, the study found that the mediating effect of eldercare satisfaction is stronger for eldercare support than for eldercare perception. This suggests that compared to the perception of facilities, planning, and services, older adults place greater importance on factors such as intergenerational support, emotional comfort, and economic security. This finding aligns with the research of other scholars, who have noted that after retirement, older adults prioritize family and social support, economic stability, and health over the physical infrastructure and services provided by community eldercare institutions. It also aligns with more general, empirically informed theories about older adults’ psychology and typical values, which likewise indicate that with advancing age, people tend to give more importance to intrinsic rather than instrumental values, and that they come to focus more strongly on the emotional aspects of their life^34,111,112. The present study shows how these tendencies are manifested in the more specific context of eldercare services, thus providing valuable guidance for government-led elder care strategies and policy frameworks.

Limitations

This study has certain limitations that also open avenues for further research. First and foremost, the research was confined to specific districts in Shanghai, which may limit the generalizability of the findings due to regional developmental disparities. Future studies should consider exploring these issues across a broader range of regions to enhance the applicability of the results. Second, the study’s concept of “community” was limited to urban environments, leaving uncertainty about the feasibility of extending, replicating, and sustaining the community-embedded elder care model in rural areas. Subsequent research could investigate the varying impacts of the embedded elder care model on the subjective wellbeing of elderly individuals in both urban and rural settings. Third, to improve survey response rates, the study focused exclusively on elderly individuals aged 60–69, which may not fully capture the wellbeing of the broader elderly population. Future research should expand this investigation to include elderly individuals over the age of 60 within community-embedded elder care models, thereby building on the findings of this study. Last but not least, the gender distribution within the sample was notably unbalanced, with female participants comprising a disproportionately larger share. While this trend aligns with demographic patterns in older adults populations, where women typically outnumber men due to longer life expectancy, it may nonetheless affect the generalizability of the findings. Gender differences have been shown to influence various aspects of aging, including care preferences, social support networks, and psychological wellbeing³⁷. As a result, the overreprensentation of female respondents may introduce gender-related bias in reported levels of eldercare satisfaction of perceived wellbeing. Future studies should strive for more gender-balanced samples or consider stratified analyses to account for potential gender-specific variations in experiences with community-embedded eldercare services.

Implications

This study contributes to the existing range of research by exploring the intrinsic relationship between community-embedded elder care and residents’ subjective wellbeing. It offers valuable perspectives on improving elder care models and enhancing the happiness of older adults. The study has several important implications:

Firstly, establishing a comprehensive community-embedded elder care service system is essential. As an emerging model that integrates the strengths of traditional care approaches, community-embedded elder care is gradually gaining prominence. However, there remains a significant gap in the development of relevant policies and regulations, resulting in an incomplete service system. This is particularly evident in the absence of standardized industry norms for hardware facilities and services, though attempts are being made by some institutions. Effective government guidance, coupled with active social participation, is crucial for the establishment of a comprehensive and standardized service system.

Secondly, enhancing hardware facilities and service quality is imperative. The study identifies the improvement of hardware facilities and service quality in elder care institutions as a pivotal factor in enhancing the subjective wellbeing of older adults. Scientific planning, coupled with facility upgrades and service quality enhancements, is fundamental to raising the intrinsic quality of community-embedded elder care institutions, thereby increasing their social credibility and acceptance.

Thirdly, because older adults prefer intrinsic, and not least emotional, over instrumental values, the design and provision of community-embedded elder care should pay special attention to affordances for emotional satisfaction. There is large potential in elder care that is genuinely community-embedded, but this potential is not sufficiently realized if the community is mainly treated as a social system or context that provides older adults with things that are instrumental to their wellbeing, and not as a personally meaningful factor in their wellbeing. This does not tell against the second observation, viz. the importance of hardware facilities and service quality. But it should be kept in mind that these are the material and functional prerequisites for living an emotionally satisfactory life, but not its content; nor are they sufficient for it.

Finally, promoting awareness and changing perceptions is critical for the sustained success of community-embedded elder care. The study underscores the significant impact of external support on the wellbeing of elderly individuals within these institutions. By increasing the level of support from families, communities, and society, the subjective wellbeing of older adults can be significantly improved. Therefore, it is essential to promote the concept of community-embedded elder care and to challenge and change existing perceptions of elder care. Achieving this requires a concerted effort from all stakeholders, including families, communities, and society at large, to ensure sustainable development in elder care practices.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Puxiang Ren: Conceptualization, methodology, data collection, formal analysis, and writing—original draft preparation. Søren Harnow Klausen: Supervision, validation, and writing—review and editing. All authors have read and agreed to the published version of the manuscript.

Funding

This study was supported by the project titled ‘Comparative study on the Innovation of Community Eldercare Service models in Megacities,’ funded by Shanghai University of Political Science and Law, under project number 2024XYB11.

Data availability

Data is provided within supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

The study was approved by the Research Ethics Committee of Shanghai University of Political Science and Law and conducted in full compliance with its guidelines and regulations. It adhered to the ethical principles outlined in the Declaration of Helsinki (1964) and its subsequent amendments.

Informed consent

Informed consent was obtained from all individual participants included in the study.