Leisure engagement in older age is related to objective and subjective experiences of aging

The study investigates which types of leisure activities in older adulthood are associated with subsequent objective and subjective experiences of aging across multiple health domains. Leisure, defined as voluntary non-work activities pursued for pleasure (e.g., hobbies, arts, sports, education, volunteering, community groups), is linked to health across the life course but is often undervalued and inequitably accessible. Prior research frequently examines single leisure domains or specific outcomes (e.g., depression, dementia, wellbeing), leaving uncertainty about comparative effects of different leisure domains on broad aging experiences. The authors propose that leisure may mitigate age-related decline via psychological, biological, social, and behavioral mechanisms, with different activities stimulating distinct mechanisms based on their “active ingredients.” Given challenges in later life (loss of social ties, living alone, low income, chronic disease) and the transition to retirement, understanding broad benefits of leisure is especially pertinent and has implications for healthcare (e.g., social prescribing). Using an outcome-wide approach, the study assesses 23 aging outcomes across seven domains and compares four leisure domains (physical, creative, cognitive, community) to determine independent associations with aging eight years later.

Existing evidence links leisure engagement to physical and mental health throughout life, including benefits of play in childhood and community engagement in reducing cognitive decline and dementia. However, many studies focus on single leisure domains (e.g., arts, physical activity) or limited health outcomes (depression, dementia, wellbeing). Some smaller, non-representative studies have examined broader aging experiences, but generalizability is limited. Theoretically, leisure’s benefits may occur through psychological (self-efficacy), biological (stress hormone reduction), social (support), and behavioral (motivation for healthy behaviors) pathways, with different activities offering distinct active ingredients (e.g., social contact, goal-directedness, sedentary/active nature). There is strong evidence for physical activity’s health benefits; evidence for creative activities is mixed, with some studies showing additional benefits in older adults and others finding less clear associations. Community and cognitive activities have been linked to outcomes such as self-rated health, frailty, disability, chronic pain, dementia, depression, and wellbeing, though effects may be stronger for mental than physical health. Gaps remain in comparing multiple leisure domains simultaneously against a broad set of objective and subjective aging outcomes in representative cohorts.

Design and data source: Outcome-wide longitudinal analysis using the U.S. Health and Retirement Study (HRS), a nationally representative cohort of adults aged 50+. Baseline for this study was the Psychosocial and Lifestyle Questionnaire administered to rotating subsamples in 2008 or 2010; follow-up occurred eight years later (2016/2018). Ethical approvals were obtained from the University of Florida and UCL; informed consent was obtained. Sample: Of 15,405 baseline psychosocial participants (2008/2010), 10,215 were re-interviewed at follow-up; 8,893 had complete leisure data; 8,771 also participated in the prior wave (for health behavior covariates) and formed the primary analytical sample. Subsamples for certain outcomes were smaller due to enhanced physical measures and age ≥65 requirements (e.g., n=7,940 with physical assessments; n=4,643 aged ≥65; n=4,131 with both). Exposure: Leisure engagement measured via the HRS Social Engagement scale (15 items used; frequency 0=never to 6=daily). Activities grouped into four domains: physical (sport/exercise, walking), creative (gardening, baking/cooking, needlework, hobbies), cognitive (reading, word games, cards/other games, writing), community (volunteering/charity work, educational courses, sports/social clubs, non-religious organizations). Domain scores were the average frequency within each domain. Outcomes: Twenty-three experiences of aging measured at baseline and 8-year follow-up across seven domains: - Daily functioning: counts of difficulties in ADLs (0–5), IADLs (0–5), and mobility (0–4+). - Physical fitness: self-reported counts of no-problem activities for strength (0–3), gross motor function (0–4), fine motor function (0–3); self-reported falls (yes/no, age ≥65); objective measures including peak expiratory flow (spirometry), grip strength (dynamometer), static balance (stance completion levels), and gait speed (98.5-inch timed walk; higher=faster; age ≥65). - Heart health: systolic and diastolic blood pressure categories and pulse measured with automated BP monitor (models adjusted for BP medication). - Weight: BMI (continuous and CDC categories: underweight/healthy <25; overweight 25–<30; obesity ≥30) and waist circumference (tape at navel). - Sleep: regular prescription sleep medication use (yes/no); rested on waking (most/sometimes/rarely-never). - Long-term physical health problems: number of chronic conditions (0–5+; hypertension, diabetes, cancer, lung disease, heart disease, stroke, psychiatric problems, arthritis); persistent pain (none/mild/moderate/severe). - Subjective perceptions of health: eyesight (with correction), hearing (excellent to poor), perceived difficulty with balance (never to often). Covariates: Demographic (age, gender, marital status, race/ethnicity), socioeconomic (education, employment, pension, household income, assets, household size), neighborhood (safety, physical disorder, social cohesion), and baseline measure of each outcome. Heart health models also adjusted for BP medication. Statistical analysis: For each outcome, regression type matched outcome scale (negative binomial for counts, linear for continuous, logistic for binary, ordered logistic for ordinal). All four leisure domains were included simultaneously. Models were adjusted for covariates and baseline outcome. Complex design and attrition were addressed using HRS sampling weights (psychosocial or physical measures weights as appropriate). Missing data were handled via multiple imputation by chained equations (20 datasets; appropriate models per variable type; max missing 16%; weights included; separate imputation by subsample). Software: Stata 17. Sensitivity analyses: (1) E-values to assess robustness to unmeasured confounding; (2) additional adjustment for baseline health covariates (cognition, depressive symptoms, prescription medication, psychiatric problems, self-rated health) and prior-wave health behaviors (alcohol, smoking); (3) models omitting baseline outcome to assess potential bias; (4) categorizing engagement frequency (none, monthly, weekly) to explore dose–response; (5) restricting to participants without chronic conditions at baseline to address reverse causation.

Sample characteristics: n=8,771 (55% female; mean age 63.18, SD 8.45; 85% White, 10% Black/African American, 5% Other; 71% married; 48% retired). Average engagement frequency (0–6): physical 3.43 (SD 1.76), cognitive 2.86 (1.16), creative 2.59 (1.16), community 1.26 (0.86). Correlations across domains r=0.21–0.34. Physical activities: More frequent physical activity predicted better outcomes eight years later across daily functioning, physical fitness, long-term physical health problems, and subjective health perceptions. Strongest associations included higher odds of good static balance (OR 1.13, 95% CI 1.09–1.17), lower perceived balance difficulties (OR 0.91, 95% CI 0.89–0.94), lower odds of mobility difficulties (OR 0.88, 95% CI 0.86–0.91), fewer ADL difficulties (IRR 0.91, 95% CI 0.87–0.95), and lower odds of having chronic health conditions (OR 0.90, 95% CI 0.88–0.93). Associations with weight and sleep were mixed (e.g., lower odds of being overweight but not waist circumference). No consistent longitudinal associations with heart health outcomes. Creative activities: More frequent engagement predicted fewer ADL difficulties (IRR 0.89, 95% CI 0.84–0.95), fewer IADL difficulties (OR 0.90, 95% CI 0.84–0.96), higher odds of good static balance (OR 1.09, 95% CI 1.03–1.16), and lower odds of sleep medication use (OR 0.84, 95% CI 0.78–0.91). No longitudinal associations with long-term physical health problems, heart health, or weight. Cognitive activities: Limited longitudinal associations; lower odds of not feeling rested after sleep (OR 0.93, 95% CI 0.88–0.98) and lower odds of rating eyesight as poor (OR 0.94, 95% CI 0.90–0.99). Community activities: No protective longitudinal associations; more frequent community engagement was associated with higher odds of chronic health conditions eight years later (OR 1.09, 95% CI 1.03–1.17). Concurrent (cross-sectional) findings: Physical activities positively associated with nearly all aspects of aging; creative activities with most domains; cognitive activities mainly with subjective eyesight but worse diastolic BP, BMI, waist circumference; community activities associated with lower odds of high systolic BP, not feeling rested after sleep, poor eyesight, and perceived balance difficulties. Sensitivity analyses: E-values for longitudinal associations ranged 1.08–1.67, indicating modest susceptibility to unmeasured confounding. After adjusting for additional health and behavior covariates, associations for physical and creative activities with daily functioning and physical fitness remained similar; other associations attenuated. Omitting baseline outcome yielded similar or slightly stronger associations for physical, creative, and cognitive activities. Dose-frequency analysis showed physical activity effects primarily for weekly engagement, whereas creative activities showed benefits at both monthly and weekly levels (less dose–response). Restricting to those without baseline chronic conditions reduced sample sizes and statistical power; estimates were directionally consistent with main analyses.

The study demonstrates that, in a nationally representative cohort of U.S. older adults, physical activity is most consistently and broadly associated with more favorable objective and subjective aging experiences over eight years, independent of other leisure domains and a wide range of demographic, socioeconomic, and neighborhood covariates and baseline outcomes. Creative activities also predict improvements, particularly in daily functioning, physical fitness, and sleep, suggesting mechanisms beyond social contact (e.g., cognitive flexibility, stress reduction, meaning and purpose). Cognitive and community activities show fewer longitudinal associations with the physical-health-oriented outcomes examined; benefits may be more pronounced for mental health or require sustained, consistent engagement. The relative lack of longitudinal associations for heart health measures and some weight metrics, despite cross-sectional links, may reflect inclusion of low-intensity activities (e.g., walking) within the physical activity index or the difficulty of modifying physiological endpoints over long periods without higher-intensity activity. Findings imply that leisure engagement may slow age-related declines in functional domains, supporting independence and potentially reducing healthcare utilization. Sensitivity analyses suggest results are reasonably robust, though modest unmeasured confounding could explain some effects. The outcome-wide approach enables comparison across multiple outcomes, clarifying where leisure engagement may exert the greatest influence (daily functioning, physical fitness, subjective perceptions of health).

Physical and creative leisure activities are associated with more positive aging trajectories across multiple domains, especially daily functioning, physical fitness, and subjective health perceptions, over an eight-year period in older adults. Encouraging and supporting regular physical activity and accessible creative engagement may help reduce age-related decline and maintain independence, aligning with policy initiatives like social prescribing. Future research should: disaggregate types and intensities of physical activity; evaluate sustained and changing engagement over longer follow-up; address potential unmeasured confounding (e.g., urbanicity, diet, social support); examine differential effects across sociodemographic groups; and improve measurement of structural inequities and culturally meaningful access to leisure. Such work can guide equitable, effective interventions to promote healthy aging.

Key limitations include: leisure engagement measured only at baseline, precluding assessment of sustained or changing participation; relatively short follow-up (8 years) given the onset of age-related changes; susceptibility to unmeasured confounding (e.g., urbanicity, diet, social support) despite extensive covariate adjustment and E-value sensitivity analyses; main models did not adjust for health and health behavior covariates that may act as mediators (sensitivity analyses may be conservative); potential bias from adjusting for baseline outcomes (addressed in sensitivity analyses); constrained measurement of gender (binary) and aggregation of several races/ethnicities into an “Other” category, limiting investigation of identity-specific effects and structural inequities; possible prodromal effects where declining health reduces engagement; smaller subsamples for some outcomes (enhanced physical measures, age ≥65) limiting power; and potential masking of associations for heart health by inclusion of low-intensity activities within the physical activity index.