Digital empowerment: Can the development of the smart health and elderly care enhance older adults’ subjective well-being?

Abstract

Objectives

The development of the smart health and elderly care has profoundly changed older adults’ daily lives and health outcomes. However, the impact of the smart health and elderly care on older adults’ subjective well-being remain unclear. This study aims to analyze the impact of smart health and elderly care on older adults’ subjective well-being and its inequality and further explore the potential mechanisms and heterogeneity.

Methods

This study employs the smart health and elderly care policy as a quasi-natural experiment, using the China Family Panel Studies panel data (N = 23,878) and a staggered difference-in-differences method to investigate the causal impact of smart health and elderly care development on older adults’ subjective well-being and its inequality.

Results

As expected, the smart health and elderly care policy significantly improves older adults’ self-rated health (B = 0.086, 95% CI 0.037–0.135, p < 0.01) and life satisfaction (B = 0.058, 95% CI 0.022–0.094, p < 0.01) while reducing depressive symptoms (B = −0.538, 95% CI −0.715 to −0.362, p < 0.01) among older adults. Furthermore, the policy helps alleviate inequalities in self-rated health (B = −0.012, 95% CI −0.021 to −0.003, p < 0.05), life satisfaction (B = −0.011, 95% CI −0.017 to −0.005, p < 0.01), and depressive symptoms (B = −0.018, 95% CI −0.025 to −0.012, p < 0.01). Mechanism analysis further suggests that intergenerational support, digital literacy, and health habits play vital mediating roles in enhancing subjective well-being and alleviating its inequality among older adults. Heterogeneity analysis reveals that the policy's positive effects are more pronounced among men, rural residents, middle- to high-income households, and older adults in the western region.

Conclusion

Our findings provide empirical evidence that developing the smart health and elderly care can enhance older adults’ subjective well-being. Consequently, the effective utilization of digital tools could be a focus of policymakers aiming to achieve healthy aging objectives.

Introduction

Promoting healthy aging has emerged as a critical key global strategy to tackling demographic aging challenges. In 2020, the 73rd World Health Assembly launched the Decade of Healthy Aging initiative. Spearheaded by the World Health Organization and endorsed by United Nations member states, this global effort seeks to promote collective actions aimed at enhancing the health, well-being, and social participation of older persons while mitigating the adverse impacts of aging. ¹ In recent years, as China progresses toward high-quality economic and social development, the country is rapidly entering a stage of moderate aging. This means that an aging population will remain a fundamental national condition for the foreseeable future. ² Against this backdrop, the needs of older adults are shifting from basic survival security to higher-level developmental pursuits. ³ Enhancing the well-being of older adults and achieving improvements in both their quality of life and longevity have become critical issues. Addressing these issues requires collective efforts from society, as they directly influence the effectiveness of China's healthy aging initiatives.

Meanwhile, the latest wave of technological advancements and industrial transformation, particularly in artificial intelligence and robotics, has opened new avenues for innovation in the health and elderly care sector. The smart health and elderly care integrates intelligent products, information system platforms, and next-generation information technologies. ⁴ This integration is driving the traditional elderly care industry toward a more intelligent and personalized transformation. To align with this trend, the Chinese government has vigorously promoted the smart health and elderly care. In 2017, the Ministry of Industry and Information Technology (MIIT) spearheaded the release of the Smart Health and Elderly Care Industry Development Action Plan (2017–2020) as the pilot policy. To implement this policy, the MIIT and two other ministries immediately launched the first batch of pilot demonstrations for smart health and elderly care in December of the same year. This initiative marked the beginning of exploring a digitally enabled model for healthy aging.

As a central-local interactive policy experiment, the smart health and elderly care pilot has become a key initiative for optimizing regional elderly care service systems and advancing a healthy aging society. Currently, the smart health and elderly care has transitioned from its initial exploratory phase to a stage of steady growth and maturity, ⁵ making significant strides in accelerating technological innovation,^6,7 establishing health information-sharing platforms,^8,9 and increasing policy support. ¹⁰ Beyond these advancements, the pilot has also improved health management among older adults^4,8 and strengthened intergenerational interactions between older adults and their children, ¹¹ potentially influencing their overall health and well-being.

While previous studies have touched upon the impact of the smart health and elderly care on older adults’ development, most remain qualitative in nature, and there is still a lack of quantitative research on their effects on subjective well-being. Key gaps persist in terms of research perspectives, empirical strategies, and potential mechanisms. Subjective well-being refers to an individual's own assessment of their health, life satisfaction, and mood. Subjective well-being inequality, in contrast, captures the relative deprivation (RD) or disparity in these assessments within a group. This study examines both the average treatment effect of the policy on well-being levels and its distributional effect on well-being inequality. This offers a more comprehensive evaluation of the policy's welfare and equity effects among older adults. Specifically, the study aims to achieve the following objectives: first, to empirically assess the causal effect of the smart health and elderly care policy on older adults’ subjective well-being and its inequality; second, to explore the mediating roles of intergenerational support, digital literacy, and health habits in the relationship between the development of smart health and elderly care and well-being outcomes; and third, to examine the heterogeneity of the policy's impact across key demographic and socioeconomic subgroups. Addressing these objectives will contribute to a more comprehensive understanding of the effectiveness of the policy and provide empirical insights for advancing healthy aging strategies.

This study contributes from the following perspectives: First, unlike prior studies that have not directly assessed the effect of the smart health and elderly care policy on older adults’ subjective well-being, this study conducts a systematic analysis, offering empirical evidence for the further expansion of the pilot. Second, by focusing on how a macro-level policy affects individual well-being, this study extends the evaluation of policy effects to the micro level, enriching research on the smart health and elderly care. Third, this study systematically examines the pathways through which the policy influences subjective well-being, with particular attention to intergenerational support, digital literacy, and health behaviors. Understanding these mechanisms enhances our knowledge of how policy interventions in this sector generate well-being effects.

Policy background and theoretical framework

Policy background of the smart health and elderly care

China is experiencing rapid industrialization, urbanization, and population aging, while information technology applications have yet to fully meet the growing public demand for health and elderly care services. To address this gap, the Chinese government introduced the Smart Health and Elderly Care Industry Development Action Plan (2017–2020) in 2017, initiating the exploration of a digitally enabled model for healthy aging. The plan's comprehensive strategy focuses on five key areas: developing technological products, promoting smart care services, establishing public service platforms, creating standard systems, and strengthening service networks and cybersecurity. Specifically, it involved developing smart health and elderly care products (e.g., health management wearables, portable monitors, home service robots) to provide age-friendly digital tools; promoting smart care services such as chronic disease management, home-based care, and online health consultations, supported by community training to enhance older adults’ ability to use digital technology for health management and social participation; constructing unified, standardized, and interoperable health and elderly care information-sharing platforms to integrate multisource resources and enable the intelligent interpretation and utilization of big data in health and elderly care; ensuring data interoperability by establishing smart health and elderly care information security standards; and strengthening broadband network infrastructure to build a smart health and elderly care service network covering households, communities, and institutions. The initiative aims to build demonstration zones that deliver effective daily care, health management, and social support services for older adults, ultimately forming an integrated smart care ecosystem.

In December 2017, the MIIT and two other ministries initiated the first batch of the smart health and elderly care pilot demonstrations. Rather than implementing the policy nationwide at once, the government adopted a gradual, pilot-based approach. The staggered rollout of the smart health and elderly care policy reflects a phased national strategy designed to promote the healthy aging industry. Specifically, the first batch of pilots was established in 2017, followed by subsequent batches in 2018, 2019, 2020, and 2021. This institutional feature implies that different regions received the policy intervention at different points in time. The selection of pilot areas aimed to achieve broad geographical coverage, representing China's vast and regionally diverse conditions. Each batch included regions with varying economic development levels, avoiding systematic bias toward any specific type of area. The timing of policy adoption in different regions was driven by the evolving needs of the national healthy aging strategy. The initial batch served as an exploratory phase to test the model's feasibility. Encouraged by positive early outcomes, subsequent batches were rolled out in a stepwise manner to expand coverage. Importantly, this variation in treatment timing across regions and over years creates a quasi-experimental setting. It allows us to employ a staggered difference-in-differences (DID) design to identify the causal effect of the policy.

The pilot selection process follows a rigorous procedure: applications are submitted to provincial authorities for preliminary review and recommendation, followed by central-level expert evaluation before final approval. An oversight mechanism provides policy and financial support while conducting regular evaluations. Pilot units that fail to meet standards may be removed from the program, ensuring quality control and continuous improvement in services for older adults.

In essence, the smart health and elderly care policy is designed to provide older adults with products and services encompassing daily life assistance, health management, recreational activities, and humanistic care. This integrated approach aims at promoting the intelligent transformation of health and elderly care and improving the well-being of the elderly. Under the unified guidance of the central policy, local governments have also introduced supporting measures, actively translating the policy content into concrete practices. This paper selects several pilot regions for illustration, as detailed in Supplementary Appendix.

Literature review and research hypotheses

The impact of the smart health and elderly care on older adults’ subjective well-being

The rise and widespread penetration of digital technology have brought new opportunities for the development of healthy aging. A systematic literature review and meta-analysis indicate that digital health interventions can effectively improve physical function among older adults living alone with chronic conditions. ¹² During public health crises, the adoption of telemedicine has been shown to positively influence the mental health of older populations. ¹³ A study based on Korean older adults, suggests that smart devices can help alleviate depressive symptoms. ¹⁴ Studies based on Chinese survey data have shown that the smart health and elderly care positively enhance older adults’ self-rated health, reduce depression risk, improve social adaptability, and increase life satisfaction.^15,16 Specifically, smart elderly care products such as blood lipid monitors and wearable devices provide real-time health monitoring and management, strengthening older adults’ sense of control over their own health status, reducing psychological anxiety. ¹⁷ Older adults utilizing smart elderly care technologies and products exhibit a measurably higher quality of life compared to nonusers.^18,19

The continuous expansion of smart health and elderly care pilots has improved the accessibility of related products and services for older adults, helping them to targetedly enhance their personal health management quality and achieve self-reliant aging. Given the significant potential of the smart health and elderly care to promote healthy aging through technological empowerment, it is reasonable to expect that digital technology interventions may directly enhance the subjective well-being of older adults.

Hypothesis 1: The development of smart health and elderly care has a positive incentive effect on older adults’ subjective well-being.

The impact of the smart health and elderly care on older adults’ subjective well-being inequality

The impact of digital technology interventions on the well-being of older adults may also involve changes at the relative level. International evidence suggests that services derived from digital technologies, such as remote consultations, have the potential to address imbalances in the distribution of healthcare resources and reduce regional disparities in health outcomes. ²⁰ However, other scholars argue that while digital health can enhance accessibility to healthcare, it may not significantly improve health inequalities ²¹ and could even exacerbate them. ²² In the Chinese context, emerging evidence suggests that digital technology significantly influences the reduction of physical and mental health inequalities, particularly benefiting disadvantaged groups. ²³

At the current stage, with the advancement of the smart health and elderly care pilot initiatives, whether the inequality in older adults’ subjective well-being can be ameliorated warrants further investigation. However, from a policy perspective, the smart health and elderly care policy aims to precisely identify diverse health and elderly care needs across different groups, enabling a multitiered, differentiated supply of services. By improving the integration and allocation of individual, family, community, institutional, and elderly care resources, the policy is expected to enhance service quality and efficiency, thereby playing a positive role in reducing intragroup inequalities in subjective well-being among older adults. ⁸ Overall, digital technology possesses characteristics of nonrivalry and nonexcludability. The widespread development of digital technology is increasingly meeting the needs of the most vulnerable groups, ²⁴ showing promise in reducing subjective well-being inequality among the older adult population. Therefore, we propose the following hypothesis:

Hypothesis 2: The development of smart health and elderly care helps mitigate older adults’ subjective well-being inequality.

Transmission mechanisms of the smart health and elderly care on older adults’ subjective well-being and its inequality.

Intergenerational support

The social support theory can better explain how the smart health and elderly care contributes to improving older adults’ subjective well-being. The theory posits that material, emotional, and informational resources obtained through social relationships can significantly improve physical and psychological health, enhance quality of life, and help individuals cope with stressors and adversities.^25–27 As a crucial form of social support, family intergenerational support has been demonstrated to promote elderly well-being across multiple dimensions. Studies based on Chinese data have found that under the “feedback model,” daily care from adult children improves elderly parents’ physical health and life satisfaction, while financial support is more effective. ²⁸ Evidence from rural elderly populations in China further corroborates that financial, instrumental, and emotional support from children contribute significantly to their physical and psychological well-being. ²⁹ Moreover, intergenerational support mitigates health disparities, with a stronger alleviating effect observed among rural residents. ³⁰

The smart health and elderly care has reshaped the interaction mechanisms between older adults and their children through digital platforms. ¹¹ These platforms allow children to monitor their parents’ health status in real time, enabling them to provide timely emotional and caregiving support. This reduces the burden of elderly care for children, particularly for only children living in different locations, who often face greater caregiving challenges. Moreover, younger generations tend to have a higher acceptance of smart technologies and demonstrate greater willingness to pay for the smart health and elderly care services. ³¹ We therefore contend that the smart health and elderly care potentially improves older adults’ well-being by enhancing the intergenerational support provided by their children.

Digital literacy

Digital literacy represents another key channel through which the smart health and elderly care influences older adults’ subjective well-being. On one hand, the smart health and elderly care policy has effectively lowered barriers to internet use for older adults by promoting age-friendly design of smart products and accessible adaptation of online applications. This not only attracts older adults intrinsically interested in digital skills ⁶ but more importantly, when seniors find such technologies within their capability, they show greater willingness to learn and adapt. ³² Meanwhile, community-based smart aging initiatives and technology training programs help address practical challenges faced by older users and narrow digital literacy gaps among different elderly groups.

On the other hand, resocialization theory in later life emphasizes that even in old age, individuals need to continually learn new skills to adapt to society, achieve self-actualization, and maintain health. ³³ In the digital context, improved digital literacy supports both physical health and quality of life. Mainstream research generally supports the positive role of digital technology in promoting older adults’ subjective well-being. Systematic reviews and meta-analyses have confirmed that digital health literacy positively influences health status and self-management among older adults, ³⁴ while low digital literacy has been identified as a major barrier to chronic disease management in this population. ¹² Evidence from China suggests that digitalization not only improves self-rated health but also alleviates health deprivation ²⁴ and reduces mental health disparities across socioeconomic groups. ³⁵ Nevertheless, some studies point to potential downsides. International evidence indicates that excessive internet use may lead to risks of social isolation and sleep problems. ³⁶ Screen exposure can cause changes in melatonin levels, leading to sleep disturbances and reduced cognitive performance. ³⁷ Furthermore, individuals who use social media for two hours or more per day may have twice the odds of perceived social isolation compared to those who use it for less than 30 min, potentially due to a reduction in offline experiences. ³⁸ Digital technology may also widen the “digital divide,” creating a “health benefits gap” among older adults. ³⁹ Some Chinese researchers further suggest a nonlinear relationship, where initial health benefits may decline with excessive use. ⁴⁰

While digital technology presents a dual impact, its negative effects often stem from excessive or improper use. On the whole, the internet has brought more benefits than harms to older adults’ lives. It is reasonable to expect that the smart health and elderly care, by guiding older adults toward appropriate digital technology use, can maximize the internet's benefits and ultimately foster sustained improvement in their subjective well-being.

Health habits

The development of smart health and elderly care may also enhance subjective well-being by fostering health habits among older adults. With the expansion of the smart health and elderly care pilot, its demonstration effect and influence have grown, encouraging older adults to prioritize their health and adopt healthy lifestyles. The policy facilitates personalized health management, allowing older adults to create health plans via digital platforms and participate in guided physical activities, ⁴¹ reinforcing preventive healthcare.

Digital health intervention combined with physical exercise may be an opportunity to promote more active and healthy ageing. ⁴² A systematic review focusing on adults aged 65 and above found that those who engaged in exercise through digital health interventions experienced significantly reduced fall rates and showed notable improvements in physical functions such as lower limb strength, as well as enhanced quality of life. ⁴³ In the Chinese context, researchers have observed that older adults using smart wearables demonstrated significantly higher engagement in preventive exercises. This suggests that smart devices may effectively motivate positive health behavior, thereby improving both physical and mental well-being.^44,45

Additionally, the impact of lifestyle on health inequalities has been increasing, with regular physical activity significantly mitigating accumulated health disadvantages and alleviating health inequality. ⁴⁶ The antidepressant effect of exercise is particularly pronounced among individuals with low socioeconomic status, ⁴⁷ promoting mental health equity. By leveraging technology and expanding service accessibility, the smart health and elderly care has provided older adults in rural and remote areas with the same health management opportunities as their urban counterparts, thereby promoting overall well-being improvements.

Hypothesis 3: The development of smart health and elderly care enhances older adults’ subjective well-being and reduces subjective well-being inequality by strengthening intergenerational support, enhancing digital literacy, and fostering health habits.

To sum up, this study proposes a theoretical framework for understanding how the smart health and elderly care influences subjective well-being and its inequality among older adults, as illustrated in Figure 1.

Figure 1.

The impact mechanism of the smart health and elderly care policy on older adults’ subjective well-being and its inequality.

Methods

Data

As secondary data analysis, this study utilizes data from the 2016, 2018, 2020, and 2022 waves of the China Family Panel Studies (CFPS). The CFPS is a nationally representative longitudinal survey led by the Institute of Social Science Survey at Peking University. It employs a multistage probability sampling strategy with implicit stratification and probability-proportional-to-size methods. Since its baseline in 2010, the survey has been conducted biennially, covering 25 provinces, municipalities, and autonomous regions with a target sample of approximately 16,000 households. It collects comprehensive information on all household members, including data on health status, individual and family characteristics, and social participation, thereby offering a reliable empirical foundation for assessing the effectiveness of the smart health and elderly care policy.

The first five batches of the smart health and elderly care policy were rolled out in 2017, 2018, 2019, 2020, and 2021, involving 86 demonstration bases across 20 provinces. The four waves of CFPS data span the period before and after the implementation of these pilots, making them well-suited for the research design. The analysis focuses on individuals aged 60 and above. After excluding invalid responses, such as ambiguous, missing, not applicable, refused, “don’t know” answers, or responses provided in mismatched contexts, the final sample consisted of an unbalanced panel with 23,878 observations. The sample includes 7588 individuals in 2016, 7333 in 2018, 4721 in 2020, and 4236 in 2022. Among them, 12,124 are male and 11,754 are female; 5902 are nonagricultural hukou (household registration) and 17,976 are agricultural hukou.

Variables

Dependent variables

Older adults’ subjective well-being: Subjective well-being reflects an individual's subjective perception and evaluation of their life circumstances.^48,49 Its measurement typically follows either a unidimensional or a multidimensional approach. While unidimensional measures often rely on closely related concepts such as happiness or life satisfaction,^50,51 they may not fully capture the construct's comprehensive nature. ⁴⁸ Consequently, a multidimensional framework is generally preferred, ⁵¹ often encompassing dimensions such as health, life satisfaction, and psychological function.^52–54 This approach is reflected in major social surveys such as the General Social Survey and the World Values Survey, which commonly use indicators including life satisfaction, self-rated health, and psychological well-being to measure subjective well-being.

For older adults, a population experiencing natural somatic decline, and in contrast to other age groups, the perceptions of both their physiological and psychological states are core constitutive elements of their subjective well-being. In the study of subjective well-being, good self-rated health is widely regarded as an indispensable measurement dimension. ⁵⁵ As well-being is fundamentally a state of being, a form of good or satisfactory living, ⁵⁶ the subjective perception of one's health directly influences the attainment of this state, with its weight being particularly pronounced among older adults. ⁵⁷ We conceptualize self-rated health as a perceived functional dimension of well-being. Similarly, we treat depressive symptoms as capturing the affective dimension of well-being. Together with life satisfaction, which represents the cognitive-evaluative dimension, these three facets provide a comprehensive, tripartite operationalization of elderly well-being. This model has been employed in recent scholarly work and represents a well-established approach in gerontological research.^58,59,60

Based on a comprehensive understanding of subjective well-being and considering scientific rigor, systematic measurement, and data availability. This study measures the older adults’ subjective well-being from three dimensions: two positive dimensions of self-rated health and life satisfaction, and one negative dimension of the depression index. First, self-rated health is older adults’ self-assessment of their physical health status, with higher values indicating better perceived health. The scale ranges from 1 to 5, where 1 represents “poor health” and 5 represents “excellent health.” Second, life satisfaction is older adults’ overall cognitive of their own quality of life, with higher values indicating greater satisfaction. The scale ranges from 1 to 5, where 1 represents “very dissatisfied” and 5 represents “very satisfied.” Third, the CES-D depression index is older adults’ self-perceived level of negative emotions. The index comprises eight questions: “I feel downhearted,” “I find it difficult to do anything,” “I have trouble sleeping,” “I feel happy,” “I feel lonely,” “I enjoy life,” “I feel sad,” and “I feel that life is meaningless.” Each item is scored on a scale of 0–3: 0 = rarely (less than one day), sometimes (1–2 days) = 1, often (3–4 days) = 2, and most of the time (5–7 days) = 3. The total CES-D score is obtained by summing all item scores, with higher scores indicating more severe depressive symptoms.

Older adults’ Subjective well-being inequality: The concept of RD originates from individuals’ perceived disadvantage compared to others, capturing within-group disparities and serving as a useful metric for micro-level inequality assessment. Initially applied in income inequality and poverty measurement, ⁶¹ RD has since been extended to fields such as health inequality.^62,63 In this study, based on RD theory, older adults with lower subjective well-being experience greater deprivation, leading to higher subjective well-being inequality. Aggregate measures such as the Gini coefficient or Theil index may fail to capture an individual's relative position in the subjective well-being distribution, whereas the Kakwani index reflects an individual's perceived disadvantage relative to others in the reference group from a micro-level perspective. ⁶¹ This makes it particularly suitable for analyzing how the smart health and elderly care policy interventions alleviate within-group disparities in subjective well-being among older adults. Following established measurement methods,^24,64 this study constructs a RD index for subjective well-being based as a proxy variable based on the Kakwani index. The calculation process is as follows: assume that there are n individuals in group W, and their subjective well-being levels are arranged in ascending order as W = (w₁，w₂，…，w _n ), where w₁ $\le$ w₂ $\le$ … $\le$ w _n . Consequently, the RD index of subjective well-being for individual i compared to individual j can be expressed as:

$$RD(w_j,w_i)=\{\begin{array}{cc}{w}_j-w_i& if{w}_j>w_i\\ 0& if{w}_j⩽w_i\end{array}$$ (1)

Subsequently, we compare the individual i's subjective well-being with all other individuals, and get the average RD of individual i's subjective well-being as follows:

$$RD(w_i)={\displaystyle \frac{1}{n{\mu }_w}(n_{w_i}^{+}\times {\mu }_{w_i}^{+}-n_{w_i}^{+}\times w_i)={\displaystyle \frac{1}{{\mu }_w}{\gamma }_{w_i}^{+}({\mu }_{w_i}^{+}-w_i)}}$$ (2)

where $w_i$ represents the subjective well-being level of individual i, ${\mu }_w$ is the average subjective well-being of all individuals, $n_{w_i}^{+}$ is the number of individuals whose subjective well-being exceeds that of individual i, ${\mu }_{w_i}^{+}$ is the average subjective well-being of the $n_{w_i}^{+}$ individuals who have better subjective well-being than i, ${\gamma }_{w_i}^{+}$ is the percentage of individuals whose subjective well-being exceeds $w_i$ . The RD index ranges from 0 to 1, where higher values indicate greater subjective well-being inequality. To facilitate understanding, a numerical example illustrating the computation of this index is also provided in the Supplementary Appendix.

Independent variables

The smart health and elderly care policy is measured using an interaction term between policy implementation timing $Pos{t}_{jt}$ and policy execution status $Trea{t}_j$ , where $Trea{t}_j$ is a binary variable, equal to 1 if individual i resides in a region with the smart health and elderly care pilot, and 0 otherwise, $Pos{t}_{jt}$ is a time dummy variable, assigned a value of 0 before the establishment of the policy, and 1 from the year of the policy implementation onward.

Control variables

To overcome the estimation bias caused by omitted variables, we incorporate a set of control factors in our analysis,^24,44 including gender (male = 1, female = 0), age, education level (years), marital status (married = 1, unmarried = 0), hukou (nonagricultural = 1, agricultural = 0), income level (scale ranges from 1 for “very low” to 5 “very high”), and social status (scale ranges from 1 for “very low” to 5 “very high”).

Mechanism variables

Based on the theoretical analyses, we suggest that the smart health and elderly care policy may have an impact on older adults’ subjective well-being and its inequality through by intergenerational support, digital literacy, and health habits. Intergenerational support consists of two aspects. The first is the proportion of healthcare expenditures, calculated as the ratio of “total healthcare expenditure over the past 12 month” to “total household expenditure over the past 12 months” from the CFPS. The second is the frequency of contact with children, measured by the question, “In the past six months, how often have you communicated with your children via phone, text, letters, or email?” The scale ranges from 0 to 6 (never = 0, almost daily = 6). Digital Literacy is proxied by whether an individual uses the internet (yes = 1, no = 0). Health habits are measured by whether an individual exercises (yes = 1, no = 0).

Descriptive statistics of the key variables are presented in Table 1.

Table 1.

Descriptive statistics for main variables.

Variables	Definition	N	Mean	SD	Min	Max
Dependent variables
Self-rated health	1–5: Poor health = 1; Excellent health = 5	23,878	2.54	1.23	1.00	5.00
Life satisfaction	1–5: very dissatisfied = 1; very satisfied = 5	23,878	4.12	0.96	1.00	5.00
CES-D	0–24: Healthy = 0; Depressive = 24	23,878	5.68	4.51	0.00	24.00
Self-rated health inequality	0–1: Relative self-rated health deprivation	23,878	0.27	0.23	0.00	0.62
Life satisfaction inequality	0–1: Relative life satisfaction deprivation	23,878	0.12	0.15	0.00	0.77
CES-D inequality	0–1: Relative CES-D deprivation	23,878	0.14	0.16	0.00	1.00
Independent variable
DID	Treatj × Postjt	23,878	0.53	0.50	0.00	1.00
Control variables
Age	Year – Birth Year	23,878	68.75	6.30	60.00	99.00
Gender	Male = 1; Female = 0	23,878	0.51	0.50	0.00	1.00
Marital status	Married = 1; Unmarried = 0	23,878	0.83	0.38	0.00	1.00
Hukou	Nonagricultural = 1; Agricultural = 0	23,878	0.25	0.43	0.00	1.00
Education	Years of Education	23,878	4.59	5.59	0.00	19.00
Income level	1–5: Very low = 1; Very high = 5	23,878	2.83	1.22	1.00	5.00
Social status	1–5: Very low = 1; Very high = 5	23,878	3.32	1.14	1.00	5.00

Model setting

Difference-in-differences model

To examine the causal effect of the smart health and elderly care policy on older adults’ subjective well-being and its inequality, this study employs a staggered DID model. The empirical specification is as follows:

$$Y_{ijt}={\alpha }_0+{\alpha }_{1}di{d}_{jt}+{\alpha }_2{X}_{ijt}+{\mu }_t+{\sigma }_j+{\epsilon }_{ijt}$$ (3)

where i denotes the individual, j represents the region, and t indicates the year. $Y_{ijt}$ is the subjective well-being or subjective well-being inequality of individual i in region j at time t. $di{d}_{jt}$ is the interaction items between pilot region $Trea{t}_j$ and pilot implementation time $Pos{t}_{jt}$ , where 1 indicates that region j implemented the smart health and elderly care policy in year t and later, and 0 otherwise. And its coefficient ${\alpha }_1$ measures the ATT impact of the smart health and elderly care policy on old adults’ subject well-being. ${\alpha }_0$ denotes a constant term, and ${\alpha }_2$ is the estimated coefficient of a set of control variables. $X_{ijt}$ represents the set of control variables including age, gender, marital status, hukou, years of education, income level, and social status. ${\sigma }_j$ denotes region fixed effect, ${\mu }_t$ denotes time fixed effect, and ${\epsilon }_{ijt}$ denotes a random error term.

Mediating effect model

To further investigate the mechanisms through which the smart health and elderly care policy influences older adults’ subjective well-being and its inequality, this study adopts a mediating effect model. ⁶⁵ This approach involves a three-step regression procedure: first, estimating the total effect of the independent variable on the dependent variable; second, examining the effect of the independent variable on the mediating variables; and third, including both the independent variable and the mediating variables in the regression model to estimate their effects on the dependent variable. Furthermore, the Bootstrap method is employed to test the robustness of the mediating effects. The model specifications are as follows:

$$Y_{ijt}={\alpha }_0+{\alpha }_{1}di{d}_{jt}+{\alpha }_2{X}_{ijt}+{\mu }_t+{\sigma }_j+{\epsilon }_{ijt}$$ (4)

$$M_{ijt}={\beta }_0+{\beta }_{1}di{d}_{jt}+{\beta }_2{X}_{ijt}+{\mu }_t+{\sigma }_j+{\epsilon }_{ijt}$$ (5)

$$Y_{ijt}={\gamma }_0+{\gamma }_{1}di{d}_{jt}+{\gamma }_2{M}_{ijt}+{\gamma }_3{X}_{ijt}+{\mu }_t+{\sigma }_j+{\epsilon }_{ijt}$$ (6)

where $M_{ijt}$ represents the mediating variables of intergenerational support, digital literacy, and health habits. In equation (4), the coefficient ${\alpha }_1$ reflects the total effect of the smart health and elderly care policy on older adults’ subjective well-being. In equation (5), the coefficient ${\beta }_1$ captures the effect of the smart health and elderly care policy on the mediating variables (intergenerational support, digital literacy, and health habits). In equation (6), the coefficient ${\gamma }_1$ represents the direct effect of the smart health and elderly care policy on older adults’ subjective well-being, while the product of coefficient ${\gamma }_2$ and coefficient ${\beta }_1$ in equation (5), denoted as ${\beta }_1{\gamma }_2$ , reflects the mediating effect of intergenerational support, digital literacy, and health habits. ${\beta }_0$ and ${\gamma }_0$ denote constant terms, while ${\beta }_2$ and ${\gamma }_3$ are the estimated coefficients of a set of control variables. All other variables defined consistently with equation (3).

Results

Parallel trend test

To ensure the validity and robustness of the DID model in evaluating the policy effect, it is essential to satisfy the parallel trend assumption. This assumption requires that before the implementation of the smart health and elderly care policy, the trends in subjective well-being between treatment and control regions should be statistically indistinguishable. To test this assumption, the interaction terms between the treatment indicator and time dummies for each survey wave are included in the regression model. The coefficients of these interaction terms are then estimated, and the results are presented in Figure 2 (due to space limitations, only the results for self-rated health are displayed, and results for other outcomes are provided in the Supplementary Appendix).

Figure 2.

Parallel trend test results.

The results indicate that before the policy implementation, the confidence intervals of the estimated interaction terms contain zero, suggesting that there were no significant differences in the growth trends of self-rated health between treatment and control regions. This establishes a reliable counterfactual: in the absence of the policy, the treatment group's trajectory would have likely mirrored that of the control group. However, after the policy implementation, the interaction terms become significantly positive, indicating that the pilot policy had a measurable impact on self-rated health. The stark contrast between the stable pretrends and the sharp post-policy divergence offers compelling evidence for a causal interpretation, effectively ruling out the concern that preexisting trends are driving the results.

This suggests that the smart health and elderly care policy effectively improves individual self-rated health and reduces inequality in self-rated health. Moreover, these findings confirm that the parallel trend assumption holds, supporting the validity of the DID model and justifying its use in this study. The successful validation of this assumption is crucial, as it lends strong credence to our subsequent estimates of the policy's net effect on subjective well-being and its inequality. Consequently, the model estimation results validate the appropriateness of employing the DID methodology.

Baseline regression results

Table 2 presents the baseline regression results on the impact of smart health and elderly care development on older adults’ subjective well-being and its inequality. Columns (1)–(3) show that the smart health and elderly care policy significantly improves self-rated health and life satisfaction while reducing depressive symptoms among older adults. Columns (4)–(6) indicate that the policy helps alleviate inequalities in self-rated health, life satisfaction, and depressive symptoms. In practical terms, these improvements likely translate into a tangible enhancement in the quality of daily life for older adults, characterized by a stronger sense of overall physical well-being and more active participation in social and family activities, thus facilitating a shift from a passive experience of aging to truly enjoying later life.

Table 2.

Baseline regression results.

Variables	Self-rated health	Life satisfaction	CES-D	Self-rated health inequality	Life satisfaction inequality	CES-D inequality
	(1)	(2)	(3)	(4)	(5)	(6)
DID	0.086*** (0.025)	0.058*** (0.018)	−0.538*** (0.090)	−0.012** (0.005)	−0.011*** (0.003)	−0.018*** (0.003)
Age	−0.014*** (0.001)	0.010*** (0.001)	−0.006 (0.005)	0.002*** (0.000)	−0.002*** (0.000)	−0.000 (0.000)
Gender	0.228*** (0.016)	−0.024** (0.012)	−1.003*** (0.058)	−0.043*** (0.003)	0.003 (0.002)	−0.031*** (0.002)
Marital status	−0.047** (0.022)	0.045*** (0.016)	−1.379*** (0.078)	0.005 (0.004)	−0.009*** (0.003)	−0.050*** (0.003)
Hukou	0.013 (0.019)	0.027* (0.014)	−1.086*** (0.069)	−0.017*** (0.004)	−0.007*** (0.002)	−0.034*** (0.003)
Education	0.010*** (0.002)	−0.005*** (0.001)	−0.072*** (0.005)	−0.003*** (0.000)	0.000* (0.000)	−0.002*** (0.000)
Income level	0.152*** (0.007)	0.130*** (0.005)	−0.341*** (0.027)	−0.026*** (0.001)	−0.021*** (0.001)	−0.012*** (0.001)
Social status	0.071*** (0.008)	0.223*** (0.006)	−0.296*** (0.028)	−0.009*** (0.001)	−0.035*** (0.001)	−0.009*** (0.001)
Constant	2.630*** (0.098)	2.262*** (0.072)	10.589*** (0.351)	0.280*** (0.018)	0.426*** (0.012)	0.295*** (0.013)
Time fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Region fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Adj. R2	0.061	0.177	0.106	0.056	0.153	0.091
N	23,878	23,878	23,878	23,878	23,878	23,878

Note: Robust standard errors clustered at the regional level are reported in parentheses. *p < 0.1, **p < 0.05, ***p < 0.01. The same applies to the following tables unless otherwise stated.

Overall, these findings demonstrate that the smart health and elderly care policy is effective on multiple dimensions of older adults’ subjective well-being. It enhances subjective well-being levels and mitigates well-being inequality, thereby empirically validating Hypotheses 1 and 2.

Endogeneity discussion

To address potential sample selection bias and endogeneity concerns, this study employs the propensity score matching DID (PSM-DID) method to reestimate the impact of the smart health and elderly care policy on older adults’ subjective well-being. The process is as follows: first, nearest neighbor matching is used to match individuals in regions with the smart health and elderly care policy (treatment group) to those in nonpilot regions (control group). Matching is performed based on age, gender, marital status, hukou, education level, income level, and social status as covariates. The balance test results are shown in Figure 3, which indicates that after matching, the characteristics of individuals in the treatment and control groups are more comparable. This suggests that the matching process is effective, ensuring a more balanced comparison. Self-selection bias and other endogeneity concerns in policy evaluation are effectively mitigated, making the estimated policy effects more reliable.

Figure 3.

The balance test results of the propensity score matching difference-in-difference (PSM-DID) model.

Table 3 presents the PSM-DID estimation results for the net effect of the smart health and elderly care policy on older adults’ subjective well-being. The results show that the policy significantly improves self-rated health and life satisfaction while reducing depressive symptoms. The policy also significantly reduces inequality in self-rated health, life satisfaction, and depression, with all coefficients negative and statistically significant at the 1% level.

Table 3.

Estimation results of the endogeneity test based on PSM-DID method.

Variables	Self-rated health	Life satisfaction	CES-D	Self-rated health inequality	Life satisfaction inequality	CES-D inequality
	(1)	(2)	(3)	(4)	(5)	(6)
DID	0.099*** (0.033)	0.219*** (0.016)	−0.544*** (0.169)	−0.025*** (0.009)	−0.020*** (0.003)	−0.120*** (0.004)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes
Time fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Region fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Adj. R2	0.054	0.187	0.109	0.054	0.170	0.091
N	12,973	12,973	12,973	12,973	12,973	12,973

These findings indicate that even after accounting for self-selection bias, the pilot policy significantly enhances subjective well-being and mitigates inequality, thereby supporting the robustness and validity of the study's conclusions.

Robustness checks

Replacement of core variable

Considering the diverse approaches to measuring the dependent variable, we conducted a robustness check to mitigate potential bias in the regression results that may arise from the selection of proxy variables. Specifically, we used happiness, a unidimensional measure, as an alternative dependent variable.^66,67 The results in Table 4 show that the coefficient of the DID estimator is statistically significant at the 1% level. This indicates that the policy enhances older adults’ subjective well-being and alleviates well-being inequality, which aligns with our main conclusions and further confirms the robustness of the regression analysis.

Table 4.

Regression results with replacement of the dependent variable.

Variables	Happiness	Happiness inequality
	(1)	(2)
DID	0.013*** (0.043)	−0.012*** (0.004)
Age	0.022*** (0.003)	−0.002*** (0.001)
Gender	−0.036 (0.033)	0.001 (0.003)
Marital status	0.258*** (0.044)	−0.023*** (0.004)
Hukou	0.271*** (0.040)	−0.026*** (0.003)
Education	0.009*** (0.003)	−0.001*** (0.001)
Income level	0.159*** (0.015)	−0.012*** (0.001)
Social status	0.398*** (0.162)	−0.030*** (0.001)
Constant	3.966*** (0.203)	0.451*** (0.017)
Time fixed effect	Yes	Yes
Region fixed effect	Yes	Yes
Adj. R2	0.094	0.086
N	23,878	23,878

Sensitivity analysis

Given the potential confounding effect of the COVID-19 pandemic as a major exogenous shock after 2020, we conducted a sensitivity analysis to examine its influence on the relationship between the smart health and elderly care policy and older adults’ well-being. We augmented our baseline regression model by incorporating a control variable for regional COVID-19 severity, ⁶⁸ measured as the standardized ratio of cumulative confirmed cases to the local population (see Supplementary Appendix for detailed calculation formula).

As shown in Table 5, the coefficient of the DID estimator remains statistically significant across all specifications after controlling for COVID-19 severity. This suggests that our main findings regarding the positive effect of the smart health and elderly care policy on elderly well-being are robust.

Table 5.

Sensitivity analysis results.

Variables	Self-rated health	Life satisfaction	CES-D	Self-rated health inequality	Life satisfaction inequality	CES-D inequality
	(1)	(2)	(3)	(4)	(5)	(6)
DID	0.213*** (0.053)	0.221*** (0.033)	−0.563*** (0.180)	−0.017* (0.009)	−0.011** (0.005)	−0.013** (0.006)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes
Time fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Region fixed effect	Yes	Yes	Yes	Yes	Yes	Yes
Adj. R2	0.069	0.167	0.089	0.064	0.151	0.079
N	8957	8957	8957	8957	8957	8957

Placebo test

To further rule out potential biases caused by unobservable factors, we conducted a placebo test. Specifically, we randomly generated pseudo-treatment and pseudo-control groups to simulate the impact of the smart health and elderly care policy on older adults’ subjective well-being. This random sampling procedure was repeated 1000 times, and the kernel density distribution of the estimated coefficients was constructed based on the 1000 simulated estimates. Detailed results are presented in Figure 4 (due to space limitations, only the results for self-rated health are displayed, and results for other outcomes are provided in the Supplementary Appendix). As illustrated, the randomly simulated coefficients are concentrated around zero and follow a normal distribution, while the estimated coefficient from the baseline regression lies completely outside this distribution. This aligns with the expectations of the placebo test, indicating that unobserved random factors can be ruled out and further confirming the robustness of our baseline results.

Figure 4.

Placebo test results.

Heterogeneity analysis

Given that group characteristics may influence the policy's impact on subjective well-being, this study conducts a heterogeneity analysis by examining differences across gender, urban-rural status, income levels, and regional economic development. Findings are shown in Table 6.

Table 6.

Results of heterogeneity analysis.

Variables	Self-rated health	Life satisfaction	CES-D	Self-rated health inequality	Life satisfaction inequality	CES-D inequality
	(1)	(2)	(3)	(4)	(5)	(6)
Panel A: Gender
Male	0.109*** (0.035)	0.062** (0.025)	−0.424*** (0.119)	−0.015** (0.006)	−0.011*** (0.004)	−0.012*** (0.004)
Female	0.062* (0.037)	0.052* (0.027)	−0.665*** (0.136)	−0.009 (0.007)	−0.011** (0.004)	−0.024*** (0.005)
Panel B: Residence
Urban	0.040 (0.047)	0.004 (0.036)	−0.522*** (0.168)	−0.011 (0.009)	−0.003 (0.006)	−0.014** (0.006)
Rural	0.100*** (0.030)	0.074*** (0.021)	−0.545*** (0.106)	−0.012** (0.006)	−0.013*** (0.004)	−0.019*** (0.004)
Panel C: Income level
High income	0.190*** (0.051)	0.013 (0.026)	−0.610*** (0.165)	−0.028*** (0.009)	−0.000 (0.004)	−0.018*** (0.006)
Middle income	0.064* (0.039)	0.036 (0.027)	−0.569*** (0.133)	−0.010 (0.007)	−0.008** (0.004)	−0.020*** (0.005)
Low income	0.007 (0.044)	0.146*** (0.040)	−0.411** (0.174)	0.002 (0.009)	−0.027*** (0.007)	−0.016** (0.007)
Panel D: Regions
Eastern region	0.100** (0.041)	0.064** (0.030)	−0.349** (0.141)	−0.011 (0.008)	−0.013*** (0.005)	−0.013** (0.005)
Central region	0.030 (0.060)	0.039 (0.042)	−0.553*** (0.211)	0.004 (0.011)	−0.005 (0.007)	−0.018** (0.008)
Western region	0.114*** (0.043)	0.074** (0.032)	−0.722*** (0.162)	−0.020** (0.008)	−0.014*** (0.005)	−0.025*** (0.006)
Northeastern Region	0.068 (0.114)	0.015 (0.086)	−0.702* (0.409)	−0.029 (0.021)	−0.005 (0.014)	−0.022 (0.015)

Specifically, the results in Panel A show that the policy's effect on enhancing subjective well-being and reducing inequality is more pronounced among men in the dimensions of self-rated health and life satisfaction, whereas its effect on reducing depressive symptoms and alleviating the corresponding inequality is stronger among women.

In Panel B, compared to urban areas, the policy demonstrates a greater impact on improving the subjective well-being of older adults in rural regions and mitigating inequalities in well-being.

The results in Panel C indicate significant disparities in policy effects across income groups. Older adults from middle- and high-income households show more notable improvements in self-rated health and depressive symptoms, while the positive influence on life satisfaction and the reduction of its inequality are more evident among low-income households.

In Panel D, the policy effects are most substantial and broad-based in western China, where notable improvements are seen in self-rated health, life satisfaction, and depressive symptoms, along with a marked reduction in well-being inequality. The eastern region also exhibits strong effects, particularly in enhancing life satisfaction and reducing depressive symptoms. In contrast, the impacts are relatively muted in the central region and statistically insignificant in the northeastern region.

Mechanism analysis

To further investigate the transmission mechanisms, this study employs a mediating effect model for empirical testing and applies the Bootstrap method to verify the significance of the mediating pathways. The results are presented in Tables 7 to 10. The pilot policy has a significantly positive impact on intergenerational support (proportion of healthcare expenditures, frequency of contact with children), digital literacy (whether the internet is used), and health habits (whether physical exercise is undertaken). Furthermore, after 1000 Bootstrap sampling tests, the above conclusions remain robust (see Supplementary appendix for detailed results). These findings confirm that the policy promotes intergenerational support, enhances digital literacy, and fosters health habits, all of which contribute to improving subjective well-being and reducing inequality. Therefore, the mediating analysis validates Hypothesis 3.

Table 7.

Mediating analysis results: proportion of healthcare expenditures.

Variables	Proportion of healthcare expenditures	Self-rated health	Life satisfaction	CES-D	Self-rated health inequality	Life satisfaction inequality	CES-D inequality
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
DID	0.179*** (0.027)	0.078*** (0.025)	0.487*** (0.090)	−0.038** (0.018)	−0.010** (0.005)	−0.017*** (0.003)	−0.008*** (0.003)
Proportion of healthcare expenditures		0.040*** (0.006)	0.251*** (0.021)	−0.114*** (0.004)	−0.009*** (0.001)	−0.008*** (0.001)	−0.015*** (0.001)

Table 8.

Mediating analysis results: frequency of contact with children.

Variables	Frequency of contact with children	Self-rated health	Life satisfaction	CES-D	Self-rated health inequality	Life satisfaction inequality	CES-D inequality
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
DID	0.271*** (0.024)	0.167*** (0.016)	0.193*** (0.057)	−0.295*** (0.013)	−0.010*** (0.003)	−0.009*** (0.002)	−0.033*** (0.002)
Frequency of contact with children		0.023*** (0.004)	0.100*** (0.015)	−0.031*** (0.003)	−0.004*** (0.001)	−0.004*** (0.001)	−0.005***(0.001)

Table 9.

Mediating analysis results: digital literacy.

Variables	Digital literacy	Self-rated health	Life satisfaction	CES-D	Self-rated health inequality	Life satisfaction inequality	CES-D inequality
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
DID	0.031*** (0.007)	0.127*** (0.024)	0.784*** (0.088)	−0.077*** (0.017)	−0.018*** (0.005)	−0.026*** (0.003)	−0.014*** (0.003)
Digital literacy		0.096*** (0.023)	0.846*** (0.083)	−0.408*** (0.016)	−0.027*** (0.004)	−0.029*** (0.003)	−0.015*** (0.003)

Table 10.

Mediating analysis results: health habits.

Variables	Health habits	Self-rated health	Life satisfaction	CES-D	Self-rated health inequality	Life satisfaction inequality	CES-D inequality
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
DID	0.034*** (0.010)	0.119*** (0.024)	0.730*** (0.088)	−0.078*** (0.018)	−0.017*** (0.005)	−0.024*** (0.003)	−0.013*** (0.003)
Health habits		0.131*** (0.016)	0.778*** (0.058)	−0.083*** (0.012)	−0.031*** (0.003)	−0.026*** (0.002)	−0.015*** (0.002)

Discussion

Principal findings

Based on mixed panel data from the 2016 to 2022 CFPS, this study employs a staggered DID model to empirically examine the impact of the smart health and elderly care policy on older adults’ subjective well-being and its inequality. Understanding the policy's effects is crucial for providing empirical insights into the construction of a healthy aging society. The findings indicate that the development of smart health and elderly care significantly improves older adults’ subjective well-being and alleviates well-being inequality, aligning with previous studies.^18,19

Furthermore, this study explores the underlying mechanisms through which the policy influences subjective well-being. First, the smart health and elderly care enhances family intergenerational support, enabling children to provide more financial assistance and emotional care to their parents. This, in turn, improves older adults’ subjective well-being and reduces its inequality. ¹¹ Second, the policy enhances older adults’ digital literacy, helping them bridge the digital divide and narrowing the gap. This technological empowerment not only improves older adults’ health self-management competencies but also facilitates their transition into post-retirement social reintegration, reducing feelings of loneliness and marginalization.^5,32 Finally, the smart health and elderly care improves older adults’ subjective well-being and reduces its inequality by fostering healthy habits. Through data-driven health management, it encourages participation in physical exercise. ⁴¹ This not only improves their physical well-being but also promotes a more positive mental outlook on aging.

However, there may be overlaps among these mediating pathways. For instance, enhanced digital literacy not only directly empowers older adults to access health information and services but may also facilitate technology-mediated intergenerational support, such as video calls or health data sharing via mobile applications. Similarly, the adoption of healthy habits, such as regular exercise tracked by smart devices, can be reinforced through both digital reminders and encouragement from family members. These interconnections suggest that the mechanisms may operate synergistically rather than in isolation. The potential interplay among these mechanisms warrants further investigation.

Interestingly, despite the overall advantages of policy effects, disparities in their impact across different groups need to be considered. In terms of gender differences, the policy's impact on self-rated health and life satisfaction is more pronounced among men, potentially due to their higher acceptance of using technological tools for health management.^19,69 Although women also experience improvements in self-rated health and life satisfaction, they tend to prioritize psychological health and social support, ⁷⁰ making the policy's effects more evident in reducing depressive symptoms and alleviating inequality.

Regarding urban–rural status differences, the policy is more effective in improving subjective well-being and reducing well-being inequality in rural areas. Due to the long-standing scarcity of high-quality healthcare resources, rural older adults exhibit a stronger response to the smart health and elderly care, particularly in terms of self-rated health and life satisfaction. Moreover, the expansion of health management services has helped narrow disparities in subjective well-being levels. In contrast, urban elderly populations already have relatively abundant healthcare resources and services, leading to weaker marginal effects of the policy. The policy's effects in urban areas are more pronounced in reducing depressive symptoms, while its role in alleviating subjective well-being inequality is relatively limited.

From an income level perspective, middle- and high-income households demonstrate more significant improvements in self-rated health and depressive symptoms, while low-income households show stronger gains in life satisfaction. This pattern aligns with Maslow's hierarchy of needs, as the diverse benefits of the smart health and elderly care policy cater to different need levels across groups. For middle- and high-income households, whose basic material needs are largely met, the policy primarily addresses higher-level needs for safety and health optimization, reflecting their stronger consumption willingness and greater propensity to invest in technology-driven health management. As a result, they achieve more substantial benefits in physical and mental health. In contrast, for low-income households, the policy's impact is most evident in enhanced life satisfaction, which is closely tied to fundamental well-being and social integration, indicating that the policy effectively meets their essential quality-of-life needs. However, preexisting consumption attitudes may influence their acceptance of smart health and elderly care products and services. Thus, despite differing pathways, the policy generates distinct yet valuable well-being gains for both groups: middle- and high-income households benefit more from optimized health management, while low-income households gain greater improvements in quality of life.

In terms of regional economic development, the policy's effects are most significant in the western region, with notable improvements in self-rated health, life satisfaction, and depressive symptoms, as well as reductions in inequality. The eastern region also exhibits strong policy effects, particularly in life satisfaction and depressive symptom reduction. The policy effects are relatively muted in the central region, while those in the northeastern region are not significant. This regional heterogeneity may be attributed to differences in developmental stages across regions, which shape policy effectiveness. In the western region, where the baseline level of older adults’ subjective well-being was relatively lower, the introduction of the smart health and elderly care policy yielded more pronounced marginal improvements, resulting in more comprehensive policy outcomes. Meanwhile, the well-developed supporting infrastructure and higher receptivity to new technologies in the eastern region provided strong support for policy implementation. In the central and northeastern regions, however, substantial economic transition pressures coupled with prominent population aging may have suppressed the immediate effectiveness and broader adoption of the smart health and elderly care policy.

Strengths and limitations

This study offers valuable insights for developing the smart health and elderly care strategies and improving the well-being of older adults. Specifically, this study examines not only the absolute level of older adults’ subjective well-being but also its distributional inequality. It thereby offers a more comprehensive evaluation of the pilot policy's effectiveness in enhancing overall welfare and promoting equity. Methodologically, we employ a quasi-experimental staggered DID design to identify the causal effect of the smart health and elderly care policy. A series of robustness strategies have been implemented to effectively mitigate potential endogeneity concerns. Furthermore, we systematically investigate the mediating roles of intergenerational support, digital literacy, and health habits and examine effect heterogeneity across subgroups. These analyses provide deeper insights into how the policy operates and for whom it is most effective, enriching the understanding of the relationship between the smart health and elderly care initiatives and older adults’ well-being. This study also contributes to the global healthy aging agenda, particularly the Decade of Healthy Aging. This initiative adopts an equity-oriented perspective to reduce health inequalities, and improve the lives of older people, their families, and communities. Our findings demonstrate that digitally enabled care can enhance well-being, reduce disparities, and optimize the health of all older adults. This evidence underscores the potential for integrating digital health strategies into national and global healthy aging agendas.

However, this study also has some limitations. First, while this study focuses on the positive impact of the smart health and elderly care policy, it does not examine potential risks associated with digital empowerment, such as increased social isolation or sleep disturbances among older adults. It is worth noting in subsequent research. Second, although we controlled for the COVID-19 pandemic through fixed effects and sensitivity analysis, its temporal overlap may still confound the policy effect. Third, the multidimensional measurement of subjective well-being, while comprehensive, relies on self-reported data that may be influenced by reporting biases and emotional states. Finally, data limitations prevent us from distinguishing directly affected individuals within pilot regions, and spatial spillovers to neighboring areas are possible. Future studies could overcome these constraints by designing targeted surveys and in-depth interviews to capture individual-level exposure or by applying spatial econometric methods, which would better isolate the policy's micro-mechanisms and spatial externalities.

Conclusion

The rapid development of smart health and elderly care presents an opportunity to enhance older adults’ subjective well-being. Using the smart health and elderly care policy as an entry point, this study combines theoretical analysis with empirical evidence based on 2016–2022 CFPS data and employs a DID model to assess the impact of the industry's development on subjective well-being and its underlying mechanisms. The findings demonstrate that the development of smart health and elderly care effectively enhances older adults’ subjective well-being. Specifically, it effectively enhances self-rated health and life satisfaction while reducing depression symptoms. Additionally, the policy helps alleviate inequalities in self-rated health, life satisfaction, and depressive symptoms. Mechanism analysis further reveals that the smart health and elderly care policy promotes subjective well-being by strengthening intergenerational support, enhancing digital literacy, and fostering health habits. The heterogeneity analysis shows that men, rural residents, middle- and high-income households, and individuals in the western region benefit the most from the policy.

Building upon these findings, this study proposes the following policy recommendations: First, policy design should be tailored to local conditions and implemented in a phased manner to further expand the scope of the smart health and elderly care pilot initiatives, thereby benefiting a broader elderly population. Regions should be encouraged to develop differentiated development models for the smart health and elderly care based on their unique characteristics, advancing the comprehensive construction of a healthy aging society.

Second, efforts should be made to enhance the precision of the smart health and elderly care products and services, emphasizing balanced resource allocation and targeted policy measures. Particular attention should be given to rural, remote, and low-income elderly populations. On one hand, targeted subsidies through financial assistance programs and public welfare funds can improve their ability to access these products and services. On the other hand, tailored promotion strategies should be developed, including utilizing new media platforms to raise awareness and willingness to adopt smart health and elderly care among low-income elderly, gradually guiding the evolution of their elderly care perspectives. Such targeted approaches help mitigate potential inequalities during policy expansion, effectively directing resources to those most in need and ensuring that digitally disadvantaged groups and regions benefit equitably from the smart health and elderly care development.

Furthermore, elderly oriented design of smart products should be strengthened with older adults’ needs as the guidance, while continuously expanding the supply of smart health and elderly care products, to ensure simple and easy operation. Simultaneously, regular digital technology training and knowledge lectures should be organized to motivate older adults to learn and use new technologies. This will help bridge the digital divide and enable them to access smart elderly care services with fewer barriers.