Socioeconomic inequalities in healthy ageing: functional capacity and community participation in older adults across 22 countries

Abstract

Background

Maintaining functional capacity and social participation is essential for healthy ageing, yet the association between socioeconomic status (SES) and these outcomes across diverse global contexts remains inadequately quantified.

Methods

This cross-sectional analysis used harmonised individual-level data from eight longitudinal ageing studies across 22 countries in adults aged ≥ 60 years (n = 70 189). A composite SES index was derived from education and household wealth. Associations between SES and three ICF-based outcomes (muscle strength, physical performance, and community participation) were assessed using multilevel mixed-effects logistic regression to account for the nested structure of the data.

Results

Among 70 189 participants (median age 68 years [IQR 63–74]; 50.9% female), strong inverse socioeconomic gradients were observed across all outcomes. Compared with the highest SES group, the lowest SES group had significantly increased odds of low muscle strength (adjusted odds ratio [aOR], 1.92; 95% CI, 1.81–2.03), poor physical performance (aOR, 1.78; 95% CI, 1.44–2.20), and low community participation (aOR, 3.91; 95% CI, 3.07–4.75; P < 0.001 for trend for all). The odds of co-occurring impairments were substantially elevated under low SES conditions, most notably for simultaneous impairment in all three domains (aOR, 9.78; 95% CI, 8.29–11.53; P < 0.001). Socioeconomic gradients varied by country income level, being most pronounced for muscle strength in low-income countries, physical performance in middle-income countries, and community participation in high-income countries.

Conclusions

SES is a fundamental determinant of healthy ageing, shaping outcomes across muscle strength, physical performance, and community participation. These findings highlight measurable SES disparities in functional capacity across diverse global contexts and provide a robust evidence base for equity-focused interventions to reduce ageing-related functional decline.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-026-27145-2.

Background

Aging involves functional decline, multimorbidity, frailty, and disability, thereby imposing escalating burdens on global healthcare and social security systems [1, 2]. In 2015, the World Health Organization (WHO) reframed healthy aging as a process of preserving intrinsic functional capacity to optimize well-being through environmental interaction. This definition transcends disease-centric paradigms to emphasize resilience despite comorbidities [3]. The International Classification of Functioning, Disability and Health (ICF) framework underpins contemporary healthy aging paradigms by requiring integrated evaluation of four pillars: body function and structures, activity limitations, participation, and contextual facilitators or barriers [4]. This framework translates biopsychosocial theory into measurable health outcomes. By operationalizing biopsychosocial paradigms, the ICF framework redirects clinical priorities toward functional health span extension [5]. Muscle strength is a key physiological determinant that declines by 1–3% annually after age 60. Concomitant mass loss (2–4%/year) disproportionately compromises functional resilience [6, 7]. Such functional deterioration independently predicts impaired activities of daily living (ADLs), falls, disability, and premature mortality [8]. Concurrently, robust social participation sustains health resilience, with socio-psychological isolation now recognized as a geroscience-defined pillar of aging [9, 10]. Consequently, integrated interventions targeting musculoskeletal capacity, mobility preservation, and community engagement are essential for achieving WHO’s healthy aging goals. Identifying modifiable socioeconomic and behavioral determinants of these domains offers actionable pathways for global policy innovation.

Socioeconomic status (SES) is a foundational determinants of population health disparities [11], with lower SES robustly associated with adverse health outcomes [12, 13], and these associations exhibit distinct trajectories across cohorts and national contexts [14]. The WHO’s Decade of Healthy Aging (2021–2030) has urgently called for addressing these inequities through actions targeting its four core action areas: combating ageism, creating age-friendly communities, delivering integrated person-centered care, and providing access to long-term care for those in need [15]. However, two critical evidence gaps hinder the effective implementation and monitoring of this global agenda. First, despite the conceptual shift towards defining healthy aging as the maintenance of functional capacity [16], prevalent disease-centric metrics and health information systems often fail to capture this multidimensional construct adequately. Second, comparative multinational data examining SES gradients across the ICF domains are limited. Furthermore, the role of pathways such as health behaviors remains insufficiently investigated. These gaps limit the identification of modifiable factors and actionable intervention targets tailored to diverse socioeconomic and national contexts.

To address these gaps, this multicohort study leverages harmonized individual-level data from 22 countries across the development spectrum to establish an integrated conceptual model linking multidimensional SES indicators to the interdependent outcomes of muscle strength, physical performance, and community participation. By providing evidence that bridges the measurement of functional capacity and the socioeconomic determinants of health, this study aims to advance the goals of healthy aging and reduce health inequities within and across countries.

Methods

Study design and population

This multinational cross-sectional analysis used harmonised individual-level data from eight population-based longitudinal aging studies across 22 countries within the Health and Retirement Study (HRS) family of surveys, including HRS (USA), MHAS (Mexico), ELSA (England), SHARE (Europe), CRELES (Costa Rica), JSTAR (Japan), CHARLS (China), and LASI (India). Detailed descriptions of each cohort are provided in the Supplement Methods. The HRS family cohorts share similar study designs, ensuring comparability across datasets. We selected the most recent wave for each cohort containing the necessary data on muscle strength, physical performance, and community participation. The specific time periods covered by the included datasets vary between 2012 and 2017 (as detailed in Table 1). This variation results from the staggered launch dates and differing biennial or triennial cycles of these international surveys. Importantly, the primary functional outcomes, particularly the objective physical examination data for muscle strength and physical performance, are core components collected in every wave. This high degree of longitudinal consistency across waves ensured the reliability of our data selection and allowed us to use the most contemporary harmonized baseline available for each country while maintaining rigorous cross-cohort comparability. This approach is consistent with established methodologies for global health aging research. Participants aged 60 years or older with complete data for the variables of interest were included. A detailed flow chart of the inclusion and exclusion process is provided in Figure S1 in Supplement.

Table 1.

Cohort characteristics: study waves, geographic distribution, and population demographics of included studies

Cohort	Wave	Year	country	Income level	Number of participants	Proportion of female (%)	Medium age (IQR)
HRS	wave 13	2016	USA	high income	3630	58.4	72(65–79)
MHAS	wave 3	2012	Mexico	Upper-middle income	1002	53.8	68(63–74)
ELSA	wave 6	2012	United kindom	high income	4756	52.9	69(64–75)
SHARE	wave 5	2012	Austria	high income	1802	50.9	69(64–74)
			Germany	high income	2730	48.9	69(64–75)
			Sweden	high income	2844	49.7	69(65–75)
			Netherlands	high income	2174	48.9	68(64–74)
			Spain	high income	2335	42.4	70(64–77)
			Italy	high income	1553	41.2	69(64–75)
			France	high income	2109	51.9	69(64–76)
			Denmark	high income	2124	51.4	68(64–75)
			Switzerland	high income	1507	48.1	69(64–75)
			Belgium	high income	2382	49.7	69(64–76)
			Israel	high income	416	36.5	67(63–73)
			Czech Republic	high income	2242	50.8	68(64–73)
			Luxembourg	high income	776	49.2	68 (64–74)
			Slovenia	high income	1129	49.3	69(64–76)
			Estonia	high income	1656	50.6	69(64–74)
CRELES	wave 5	2012	Costa Rican	Upper-middle income	1470	59.9	63(61–66)
JSTAR	wave 4	2012	Japan	high income	1819	48.6	69(65–74)
CHARLS	wave 3	2015	China	Upper-middle income	1677	46.0	65(62–70)
LASI	wave 1	2017	India	Lower-middle income	28,056	51.8	67(63–72)

Abbreviations: HRS the US Health and Retirement Study, MHAS the Mexican Health and Aging Study, ELSA the English Longitudinal Study on Aging, SHARE the Survey of Health, Aging and Retirement in Europe, CRELES the Costa Rican Longevity and Healthy Aging Study, JSTAR the Japanese Study of Aging and Retirement, CHARLS the China Health and Retirement Longitudinal Study, LASI the Longitudinal Aging Study in India

All included cohort studies obtained ethical approval from their respective institutional review boards, and all participants provided written informed consent (Methods in Supplement). As this secondary analysis used exclusively de-identified, anonymised data from these cohorts, ethical approval was not required for this study.

Assessment of SES

SES was assessed as an individual-level variable using two harmonized indicators: educational attainment and household wealth. Following established methodologies for cross-national aging studies [13], education was mapped into three predefined levels according to national systems: Low (≤ Primary), Medium (Secondary/Vocational), and High (≥ University). Household wealth was categorized into country-specific tertiles (T1 = low, T2 = middle, T3 = high) based on each individual’s relative position within their respective national distribution. This approach defines SES relative to the specific economic context of each nation, thereby accounting for macro-economic heterogeneity across 22 countries while capturing individual-level socioeconomic disparities. We utilized household wealth rather than individual income to better reflect the long-term socioeconomic resources available to older adults, many of whom are retired.

To calculate the composite SES index, ordinal scores of 1, 2, and 3 were assigned to the Low, Medium, and High levels of both education and household wealth, respectively. These scores were summed to create a composite score ranging from 2 to 6, which was subsequently categorized into four levels: low SES (score of 2), low-middle SES (3), middle-high SES (4), and high SES (5–6). Additionally, a 3 × 3 education-wealth matrix crossing these indicators defined nine mutually exclusive SES strata, providing a finer 1–9 continuous-like gradient to examine the dose-response relationship between individual-level SES and health outcomes.

Outcomes

Aligned with the International Classification of Functioning, Disability and Health (ICF) framework [4], this study operationalized three core domains of functional ability: (1) Muscle strength, reflecting the body functions component (ICF category b730: muscle power functions), was objectively measured using handgrip dynamometry with cohort-validated, sex-specific cut-points; (2) Physical performance, representing the activities component (ICF category d4: mobility), was evaluated through adapted protocols including balance tests, gait speed, chair-stand assessments, and validated questionnaire items on mobility limitations; (3) Community participation, corresponding to the participation component (ICF category d9: community, social and civic life), was captured via self-reported involvement in social, recreational, and civic activities such as sports clubs, volunteer work, and community events. Detailed cohort-specific validation criteria and alignment with ICF codes are provided in Table S 1 in Supplement.

Covariates

The following variables were selected as confounders based on prior literature to adjust for potential bias in the estimated associations between SES and functional outcomes: age (continuous); sex (male/female); body mass index (BMI; categorized as underweight, normal weight, overweight, obese); alcohol consumption (never/ever); smoking status (never/ever); physical activity (defined as participation in high-intensity activity at least once per week or not); and self-reported, physician-diagnosed conditions (hypertension, diabetes, cancer, lung disease, heart disease, stroke, and arthritis), then further categorized as 0, 1, 2, or ≥ 3 conditions.

Statistical analysis

Statistical analyses utilized Stata/MP (v17.0) for data harmonization and multivariable regression, and R (v4.4.1) for visualization. Descriptive statistics were summarized using means (IQR) for continuous variables and frequencies (percentages) for categorical variables. Group comparisons for continuous variables were performed using ANOVA, while chi-square tests were applied for categorical variables.

Multivariable adjusted logistic regression models were employed to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the association between SES and outcomes across different country-income groups, classified according to World Bank criteria at the time of cohort enrollment. Analyses were performed on harmonized individual-level data pooled from all participating cohorts into an integrated database. The models were stratified by country-income groups classified according to World Bank criteria at the time of cohort enrollment. The analyses examined the independent associations of educational attainment and household wealth on outcomes, followed by the association with categorized SES composite scores. Additionally, we investigated the combined associations of education and wealth through various categories and visualized the ORs using heat maps. To account for between-cohort variation and the nested structure of the data, cohort-level fixed effects were included in the multivariable models.

To examine the robustness of the findings, subgroup analyses were conducted stratified by sex, age groups, BMI categories, and number of chronic conditions. To assess the robustness of the socioeconomic gradient against potential confounding by mental health, we conducted a sensitivity analysis stratified by depressive symptoms. Depressive symptoms were harmonized across eight cohorts based on population-specific validated scales and established cut-offs (detailed in Table S1). Interaction terms (SES times depression) were tested using Wald tests to evaluate effect modification across socioeconomic strata. Sensitivity analyses involved: (1) excluding participants with major illnesses potentially affecting SES or causing severe systemic health impairment (specifically, stroke, cardiac disease, and cancer) to assess the stability of the education-wealth gradients; and (2) evaluating household income as an alternative indicator of SES to examine the consistency of associations using the same multivariable logistic regression models employed in the primary analyses.

Results

The final sample consisted of 70 189 individuals aged 60 years or older (median age 68 years [IQR 63–74]; 50.9% female). The national, sex, and age distributions across cohorts are summarized in Table 1. Among them, 15 843 (22.6%) participants were classified as low SES, 18 753 (26.7%) as lower-middle SES, 18 563 (26.4%) as middle-high SES, and 17 030 (24.3%) as high SES. Participants with lower SES were older, more likely to be female, more often unmarried, and had a higher prevalence of underweight, a lower prevalence of smoking and alcohol consumption, higher levels of physical inactivity, and a lower number of chronic diseases (all p < 0.001). In the overall population, the prevalence of low muscle strength was 14.8%, low physical performance 20.5%, and low community participation 64.8%. Each of these outcomes showed a graded inverse relationship with SES. Specifically, the prevalence of poor health outcomes significantly increased with decreasing SES levels (all p < 0.001). Detailed descriptive statistics stratified by SES category are provided in Table S 1 in Supplement.

The Venn diagram (Fig. 1) illustrates that the proportion of individuals classified as exhibiting healthy aging (defined as normal muscle strength, physical performance, and community participation) decreased significantly with decreasing SES. In the highest SES group, 49.0% of individuals were categorized as exhibiting healthy aging, whereas this proportion was only 14.1% in the lowest SES group. Crucially, a higher degree of co-occurrence of reduced muscle strength, physical performance, and community participation was observed with decreasing SES. The proportion of individuals with all three domains impaired was 1.6% in the highest SES group, but this proportion increased significantly to 8.7% in the lowest SES group. A consistent graded pattern was observed across all SES categories, with progressively higher odds associated with lower SES levels for every impairment combination (Table 2). Notably, the strength of these associations increased markedly with the number of co-occurring conditions. For isolated impairments, the low SES group showed significantly elevated odds compared to the high SES reference group, with adjusted odds ratios (aORs) of 1.64 (95% CI 1.34–1.92, p < 0.001) for low muscle strength, 1.98 (1.74–2.24, p < 0.001) for low physical performance, and 3.35 (3.14–3.57, p < 0.001) for low community participation. The odds substantially increased to aORs of 4.15 (3.27–5.27, p < 0.001) for combined low muscle strength and physical performance, 6.04 (5.31–6.87) for low muscle strength with low community participation, and 4.85 (4.39–5.36, p < 0.001) for low physical performance with low community participation. The odds were highest for co-occurrence of all three impairments (aOR 9.78, 95% CI 8.29–11.53, p < 0.001).

Fig. 1.

Socioeconomic status and the co-occurrence of low muscle strength, low physical performance, and low community participation.Venn diagrams represent the prevalence and overlap of these conditions across different socioeconomic levels: A High SES (N=17,030); B Middle-high SES (N=18,563); C Lower-middle SES (N=18,753); and D Low SES (N=15,843). The green circles indicate the proportion of individuals categorized as "Healthy aging" within each group

Table 2.

Associations between socioeconomic status and isolated and co-occurring low muscle strength, low physical performance, and low community participation

	Low muscle strength		Low physical performance		Low social participation		Low muscle strength & low physical performance		Low muscle strength & Low social participation		Low physical performance & Low social participation		Low muscle strength & low physical performance & Low social participation
	Ncase/ N total	OR (95%CI)	Ncase/ N total	OR (95%CI)	Ncase/ N total	OR (95%CI)	Ncase/ N total	OR (95%CI)	Ncase/ N total	OR (95%CI)	Ncase/ N total	OR (95%CI)	Ncase/ N total	OR (95%CI)
SES category
Low	315/2542	1.64(1.34–1.92)‡	532/2759	1.98(1.74–2.24)‡	7430/9657	3.35(3.14–3.57)‡	228/2455	4.15(3.27–5.27)‡	1455/3682	6.04(5.31–6.87)‡	2276/4503	4.85(4.39–3.36)‡	1380/3607	9.78(8.29–11.53)‡
Lower-middle	385/4235	1.34(1.16–1.54)‡	645/4495	1.50(1.33–1.68)‡	8973/12,823	2.70(2.55–2.85)‡	203/4053	2.63(2.08–3.34)‡	1361/5211	3.82(3.37–4.33)‡	2221/6071	3.17(2.89–3.49)‡	1115/4965	5.32(4.52–6.27)‡
Middle-high	452/5695	1.22(1.06–1.39)§	724/5967	1.34(120–1.49)‡	8183/13,426	1.92(1.82–2.02)‡	172/5415	1.81(1.42–2.31)‡	1011/6254	2.27(2.00–2.58)‡	1923/7166	2.26(2.05–2.48)‡	855/6098	3.46(2.93–4.08)‡
high	529/8879	1(ref)	748/9098	1(ref)	5526/13,876	1(ref)	126/8476	1(ref)	523/8873	1(ref)	964/9314	1(ref)	264/8614	1(ref)

‡Test significance p < 0.001. §Test significance p < 0.01

Further analyses using a 9-level SES category revealed differential gradients in the associations with low muscle strength, poor physical performance, and low community participation across country-income groups (Fig. 2). Consistently, within each economic stratum, odds exhibited stepwise increases with lower educational attainment; similarly, within matched education categories, odds escalated progressively with declining wealth. The SES-associated odds magnitude varied by outcome and economic context: low-income countries showed the strongest association for low muscle strength (aOR 2.23, 95% CI 1.76–2.81, p < 0.001); middle-income countries for poor physical performance (aOR 4.17, 95% CI 2.59–6.71, p < 0.001); and high-income countries for low community participation (aOR 4.73, 95% CI 4.35–5.15, p < 0.001). Cohort-specific and pooled analyses (Figure S 2–4 in Supplement) confirmed strong relations between low SES and all outcomes: low muscle strength (aOR 1.92, 95% CI 1.81–2.31; I² = 68.2%, p = 0.003), poor physical performance (aOR 1.78, 95% CI 1.44–2.20; I² = 58.8%, p = 0.018), and low community participation (aOR 3.91, 95% CI 3.07–4.75; I² = 78.2%, p < 0.001).

Fig. 2.

Associations of education-wealth combinations with odds of low muscle strength, physical performance, and community participation. Models were adjusted for demographic factors (age, sex, and marital status), lifestyle factors (BMI, smoking, alcohol consumption, and physical activity), and chronic physical conditions

Subgroup analyses further confirmed the primary findings across population strata (Figure S 5–7 in Supplement). The gradients in the associations of combined education-wealth categories with low muscle strength and low community participation were markedly steeper in individuals aged 60–70 years, those classified as underweight or normal weight, and those with ≤ 1 chronic physical condition. Conversely, the odds of poor physical performance increased more prominently among individuals aged ≥ 70 years and those with ≥ 2 chronic physical conditions.

Sensitivity analyses, excluding individuals with cancer, heart disease, or stroke (conditions potentially causing significant wealth depletion or severe systemic health impacts), demonstrated that the clear gradients for education and wealth persisted across all three adverse outcomes (Figure S8 in Supplement). The gradient associated with educational attainment remained more pronounced than that associated with wealth. Sensitivity analysis stratified by depressive symptoms (n = 68,544; Table S3 & Figure S9) demonstrated that the inverse socioeconomic gradient remained robust across all functional outcomes. Individuals in the lowest SES category (Level 9) exhibited significantly higher odds of impairment in both non-depressed and depressed subgroups (aORs: 1.91–4.77 and 1.73–3.78, respectively). Interaction tests were non-significant for all domains (P > 0.05), reinforcing SES as a fundamental determinant of healthy aging that operates independently of individuals’ baseline psychological profiles. When household income was used as an alternative indicator of SES, pooled analyses revealed significant associations with the adverse outcomes: pooled OR for low household income was 1.52 (95% CI 1.32–1.72; p = 0.019) for low muscle strength, 1.38 (95% CI 1.17–1.58; p < 0.001) for poor physical performance, and 1.71 (95% CI 1.50–1.92; p = 0.001) for low community participation (Figure S 10–12 in Supplement).

Discussion

This multinational cross-sectional study of 70 189 older adults across 22 countries demonstrates significant socioeconomic gradients in healthy aging. Lower SES, measured by education and household wealth, was robustly associated with elevated odds of impaired muscle strength, reduced physical functioning, and restricted community participation. Significant disparities in healthy aging prevalence were observed, ranging from 49.0% in high-SES to 14.1% in low-SES groups. These associations exhibited cross-cultural consistency, persisted after adjustment for diseases and lifestyle factors. Sensitivity analyses using household income as an alternative SES indicator confirmed these patterns, underscoring socioeconomic inequality as a fundamental barrier to achieving healthy aging. While prior studies established SES as a determinant of disease multimorbidity [12, 17], our findings extend this evidence by revealing SES-driven functional inequities across the ICF continuum.

The WHO advocates for lifecourse approaches in shaping health trajectories, emphasizing that aging processes commence at birth. Thus, later-life health reflects cumulative influences rather than solely geriatric factors, necessitating a focus on modifiable determinants across life stages [18, 19]. Socioeconomically advantaged individuals accrue lifelong benefits through wealth and education, higher health literacy, superior healthcare access, secure housing, and nutritional stability that collectively enhance late-life functional capacity [20, 21]. Socioeconomic health disparities emerge from multilevel determinants (individual, familial, community-level) operating differentially across developmental phases [22], making the identification of intervention targets imperative for health equity [23]. While global health has continued to improve over recent decades, progress has been uneven across and within countries, with significant disparities persisting between nations of different income levels [24]. Our analyses revealed context-dependent gradients: strongest SES-muscle strength associations in lower-middle-income countries may reflect subsistence imperatives where malnutrition, hazardous occupations, and rehabilitation deficits compromise physiological foundations [25, 26]; sharpest physical functioning disparities in middle-income nations may arise from environmental barriers and unmanaged comorbidities during rapid epidemiological transitions [27]; most pronounced SES-community participation inequities in high-income settings may stem from economic disconnectedness, opportunity stratification, and psychosocial stressors when basic needs are met [28]. Education and wealth were associated with independent and synergistic patterns. Education cultivates health knowledge, cognitive reserves, and behavioral competencies, while wealth provides material buffers against health shocks. Together, these factors enable higher health literacy, healthier behaviors, and optimized healthcare utilization among high-SES groups [29].

‘Healthy Aging’ is not about the absence of disease, but about the capacity to maintain intrinsic abilities and functional performance despite impairments or chronic conditions, enabling meaningful environmental engagement [16]. Such functionally independent aging elevates individual well-being while generating societal benefits through reduced public expenditures [19].While lifestyle factors and chronic conditions are known to influence functional decline, our analysis demonstrates that the inverse socioeconomic gradient across all three domains remains robust even after adjusting for physical activity, BMI, and disease burden.This suggests that SES acts as a fundamental determinant of healthy ageing that operates independently of these specific behavioral and clinical factors. The persistence of these disparities after adjustment emphasizes that interventions focusing solely on biomedical management or individual behaviors may be insufficient to close the health equity gap. Chronic conditions may accelerate functional decline through pathways like sarcopenia [30], WHO’s recommendation of ≥ 3 sessions/week for preserving musculoskeletal health [31]. Besides, family support may play a crucial role in maintaining older adults’ functional capacity [32]. Chronic conditions substantially compromised muscle integrity and functional capacity.

This study has several key strengths, including a large-scale, multinational sample of older adults from diverse socioeconomic and geographical settings, and a harmonised data protocol that enhances the generalisability of the findings. However, we acknowledge the following limitations. First, although environmental factors, particularly age-friendly societal infrastructure, play a critical role in functional health [3], their variability across 22 countries precluded standardised measurement, limiting our ability to assess the influence of built environments. Second, while reverse causality due to health selection is less salient in older populations than during working age, and sensitivity analyses excluded participants with major health conditions, residual bidirectional relationships between SES and health may remain [22]. Third, unmeasured confounders such as childhood SES, life-course adversity, and psychosocial stress may partly account for the observed associations [33]. Finally, the use of cross-sectional data is insufficient to adequately capture the complex causal mechanisms that unfold over the life course. Future studies should incorporate prospective cohort designs to trace temporal sequences and dynamic processes, as well as mixed-methods approaches to contextualize and triangulate longitudinal quantitative findings.

Conclusion

In conclusion, we established SES as a structural determinant of healthy aging, evidenced by profound three-dimensional disparities across musculoskeletal and physical functioning, and social engagement. These inequities underscore that achieving health equity requires redefining healthy aging beyond disease avoidance toward sustaining functional capacity and meaningful participation. Our findings robustly support WHO’s function-centered paradigm and necessitate policy pivots from biomedical management toward multisectoral interventions targeting social-behavioral determinants.

Abbreviations

ADLs

Activities of Daily Living

aOR

Adjusted Odds Ratio

BMI

Body Mass Index

CHARLS

China Health and Retirement Longitudinal Study

CI

Confidence Interval

CRELES

Costa Rican Longevity and Healthy Aging Study

ELSA

English Longitudinal Study of Ageing

HRS

Health and Retirement Study

ICF

International Classification of Functioning, Disability and Health

IQR

Interquartile Range

LASI

Longitudinal Aging Study in India

MHAS

Mexican Health and Aging Study

OR

Odds Ratio

SES

Socioeconomic Status

SHARE

Survey of Health, Ageing and Retirement in Europe

JSTAR

Japanese Study of Aging and Retirement

WHO

World Health Organization

Authors’ contributions

ZT and XYZ conceived and designed the study. ZT, JTG, and XYW were responsible for raw data cleaning and processing. ZT conducted the formal analysis, generated visualizations, and drafted the initial manuscript. XYZ, BXC, SG, YHZ, WY, and ZYR reviewed and revised the manuscript. XYZ also acquired funding and provided supervision. All authors read and approved the final manuscript.

Funding

The study was supported by the Chinese Academy of Engineering Strategic Consulting Project (Project No.2025-XZ-119) and Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (Grant No. 2022-ZHCH330-01).

Data availability

The original data for this study are derived from multiple publicly available longitudinal aging surveys. The original data can be accessed upon application through the Gateway to Global Aging Data (https://g2aging.org) or the official websites of each constituent study: the US Health and Retirement Study (HRS; https://hrsdata.isr.umich.edu/), the Mexican Health and Aging Study (MHAS; https://www.mhasweb.org), the English Longitudinal Study of Aging (ELSA; https://www.elsa-project.ac.uk/accessing-elsa-data/), the Survey of Health, Aging and Retirement in Europe (SHARE; https://share-eric.eu/), the Costa Rican Longevity and Healthy Aging Study (CRELES; https://populationsciences.berkeley.edu/creles/), the Japanese Study of Aging and Retirement (JSTAR; https://www.rieti.go.jp/en/projects/jstar), the China Health and Retirement Longitudinal Study (CHARLS; http://charls.pku.edu.cn), and the Longitudinal Aging Study in India (LASI; https://lasi-india.org/). Access to each dataset is subject to the terms and conditions of the respective studies.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki. All eight population-based longitudinal aging studies included in this analysis (HRS, MHAS, ELSA, SHARE, CRELES, JSTAR, CHARLS, and LASI) obtained ethical approval from their respective institutional review boards or ethics committees. Written informed consent was obtained from all participants prior to their inclusion in the primary studies. Detailed information regarding the specific ethics approval committees and protocols for each cohort is provided in the Supplementary Methods.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.