Friday, February 23, 2024

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

Ethical approval

This research complies with all relevant ethical regulations. All participants gave informed consent, and this study has approval from the University of Florida (IRB201901792) and the University College London Research Ethics Committee (project 18839/001). HRS offers financial payments as tokens of appreciation to respondents for participating, but these were not intended as compensation.


Participants were drawn from the Health and Retirement Study (HRS), a nationally representative study of more than 37,000 individuals over the age of 50 in the US55. The Health and Retirement Study (HRS) was initiated by the National Institute on Aging and conducted by the Institute for Social Research at the University of Michigan to track the Baby Boom generation’s transition from work to retirement. The initial HRS cohort was interviewed for the first time in 1992 and followed up every two years, with other studies and younger cohorts merged with the initial sample. Together, these studies create a fully representative sample of individuals over the age of 50 in the United States. Further details on study design are reported elsewhere55. In this study, we combined seven HRS public datasets (HRS 2018 Tracker Final Release [V1.0], RAND HRS Longitudinal File 2018 [V1]; RAND HRS Detailed Imputations File 2018 [V1]; 2008 RAND HRS Fat File [V3A]; 2010 RAND HRS Fat File [V5F]; 2016 RAND HRS Fat File [V2B]; 2018 RAND HRS Fat File [V2A]). Raw data are available from HRS ( and the RAND Center for the Study of Aging (

At each wave of HRS, a rotating random 50% subsample of participants were invited to an enhanced interview and given a leave-behind Psychosocial and Lifestyle Questionnaire to return by mail, which included questions on leisure engagement56. Participants were eligible to complete this psychosocial questionnaire in 2008 or 2010, which we have combined to form the baseline of our study. In 2008, 8296 participants were invited to complete this questionnaire, and 7073 (85%) returned it. In 2010, 11,213 were eligible, and 8332 (74%) participated. Of the 15,405 participants who participated at baseline, 10,215 also participated in the HRS core survey at our follow-up eight years later (2016/2018) and were thus eligible for inclusion in our study. Of these, 8893 participants had complete data on leisure engagement, and 8771 also participated in the previous wave (in which health behavior covariates were measured), forming our final analytical sample for outcomes measured in the core survey (Table S1). Three additional limitations reduced our sample size further for some outcomes: completion of enhanced physical assessments at follow-up (n = 7940), aged 65 and over at follow-up (n = 4643), and both restrictions combined (n = 4131).

Leisure engagement

The HRS Social Engagement scale was measured in the psychosocial questionnaire at baseline56. This scale included 18 consistent items across 2008 and 2010, three of which were excluded from this study (caring for sick or disabled adults, praying privately, using a computer for email, internet, or other tasks) as they were not typical leisure activities. This left 15 activities, which have previously been categorized into four domains: (a) physical activities (sport/exercise, walking), (b) creative activities (gardening, baking/cooking, needlework, and hobbies), (c) cognitive activities (reading, word games, cards or other games, and writing), and (d) community activities (volunteering, charity work, educational courses, sports or social clubs, non-religious organizations)28. Participants reported how frequently they engaged in each activity on a seven-point scale, from never (0) to daily (6). We created an index for each domain, averaging the frequency of engagement in all activities within that domain.

Experiences of aging

All outcomes were measured at baseline (2008/2010) and eight years later (2016/2018). The subsample in which each outcome was measured is detailed in Table S1.

Daily functioning

Included the number of difficulties with ADLs (ranging from 0 to 5; from bathing, eating, dressing, walking across a room, and getting in or out of bed), the number of difficulties with IADLs (ranging from 0 to 5; from using a telephone, taking medication, handling money, shopping, and preparing meals), and number of difficulties with mobility, from walking one block, several blocks, and across a room, jogging one mile, and reaching/extending arms up (0, 1, 2, 3, 4 or more).

Physical fitness

Included three self-reported indices, measured as the number of activities with which participants did not have problems within: strength (ranging from 0-3; stooping, kneeling or crouching, pushing or pulling a large object, lifting or carrying weights over ten pounds [like a heavy bag of groceries]), gross motor function (ranging from 0 to 4; walking one block, walking across a room, getting in or out of bed, bathing), and fine motor function (ranging from 0 to 3; picking up a dime, eating, dressing). On each of these outcomes, higher scores indicate better physical fitness. Participants aged 65 and over self-reported whether they had fallen down in the last two years (yes, no; falls). We also included four objective measures of physical fitness, assessed during the enhanced face-to-face HRS interview57. Lung function was measured with peak expiratory flow using a Mini-Wright Peak Flow Meter, taken as the average of three measures each 30 s apart. Grip strength was measured with a pistol hand grip device (Smedley spring-type hand dynamometer), taken as the average of two measures on each hand. Static balance was evaluated with three separate, progressively more difficult stances, with a variable derived to indicate completion of these stances (none, side-by-side only, semi-tandem, tandem). Gait speed was measured in participants aged 65 and over with the timed walk test (time to walk a 98.5-in. course twice), with times reversed so that higher scores indicate faster gait speed.

Heart health

Was measured with systolic and diastolic blood pressure and pulse, all assessed using an Omron HEM-780 Intellisense Automated blood pressure monitor with ComFit cuff on the participant’s left arm, taken as the average of the final two of three measurements 45–60 s apart. Using the American Heart Association guidelines, we categorized systolic blood pressure as normal (less than 120 mm Hg), elevated (120–129 mm Hg), hypertension stage one (130–139 mm Hg), or hypertension stage two (140 mm Hg and above) and diastolic blood pressure as normal or elevated (less than 80 mm Hg), hypertension stage one (80-89 mm Hg), or hypertension stage two (90 mm Hg and above). Models including these measures were adjusted for whether participants took blood pressure medication (yes, no).


Descriptors were body mass index (BMI), calculated using interviewer-recorded weight and height, and waist circumference, measured with a tape measure at the level of the participant’s navel. We categorized BMI using the Centers for Disease Control and Prevention (CDC) guidelines as people who were underweight or healthy weight (below 25), overweight (25 to <30), or people with obesity (30 and above). The underweight and healthy weight categories were combined due to the low proportion of participants who were underweight.


Measures were self-reported, including whether participants regularly take prescription medication to help them sleep (yes, no) and how often they feel really rested when they wake up in the morning (most of the time, sometimes, rarely/never).

Long-term physical health problems

Were self-reported as the number of chronic health conditions (0, 1, 2, 3, 4, 5 or more; from high blood pressure, diabetes, cancer, lung disease, heart disease, stroke, psychiatric problems, arthritis) and the degree of persistent pain experienced (none, mild, moderate, and severe).

Subjective perceptions of health

Included participants’ ratings of their eyesight using glasses or corrective lenses as needed (excellent, very good, good, fair, poor, and blind), hearing (excellent, very good, good, fair, and poor), and perceived difficulty with balance (never, rarely, sometimes, and often).


Covariates were measured in the HRS core survey at baseline (2008/2010). Demographic factors were age (years), gender (men, women), marital status (married [including cohabiting], unmarried [separated, divorced, widowed, never married]), and race/ethnicity (White [including Caucasian], Black [including African American], Other [including American Indian, Alaskan Native, Asian or Pacific Islander, Hispanic, Other]). Socioeconomic factors were educational attainment (less than high school, high school, college, postgraduate), employment status (employed, retired, not working [including unemployed, temporarily laid off, disabled, homemakers]), pension status (yes, no), total household income (US dollars), total assets (US dollars), and household size (count of other household members). Neighborhood factors were self-reported safety (excellent/good, fair/poor), physical disorder (ranging from 1 to 7; rated presence of vandalism, graffiti, rubbish, vacant, deserted houses, and crime), and social cohesion (ranging from 1 to 7; feels part of this area, trusts people, people are friendly, people will help).

Statistical analysis

Using an outcome-wide approach26,27, we tested the associations between frequency of engagement in each leisure domain (physical, creative, cognitive, and community activities) and aging experiences eight years later in regression models. The type of regression was determined by the outcome; negative binomial regression was used for count outcomes (to deal with overdispersion), linear regression for continuous, logistic regression for binary, and ordered logistic regression for ordinal outcomes. Models included all four leisure domains simultaneously and were adjusted for demographic, socioeconomic, and neighborhood covariates and the baseline measure of the outcome. All analyses of heart health were also adjusted for whether participants were taking blood pressure medication (yes, no). The cross-sectional associations between leisure engagement and each experience of aging at baseline are included in the Supplementary Materials, along with the unadjusted models.

We accounted for the complex survey design and attrition using probability weights provided by HRS, including either the weight for the psychosocial questionnaire subsample or the physical measures subsample, dependent on how each outcome was measured. For participants with missing data on aging experiences outcomes or covariates, we imputed data using multiple imputations by chained equations (MICE)58. We used Poisson, multinomial logistic, ordered logistic, and logistic regression according to variable type, generating 20 imputed data sets (maximum missing data 16%; Table S2). The imputation model included all variables used in analyses and sampling weights. Separate imputation models were run for each subsample (1) core survey, 2) physical measures, 3) aged 65+, and 4) physical measures and aged 65+; Table S1). All analyses were performed using Stata 1759.

In sensitivity analyses, we computed E-values as indicators of how robust findings were to potential unmeasured confounding29, using the Stata evalue package60. We also performed four additional sensitivity analyses. First, we additionally adjusted analyses for health covariates (cognition, depressive symptoms, prescription medication, psychiatric problems, self-rated health measured at baseline) and health behavior covariates (alcohol use and smoking measured at the wave prior to baseline). Second, as there were concerns about potential bias due to controlling for the outcome at baseline, we also repeated the longitudinal adjusted analyses after omitting the baseline outcome measure. Third, we included different levels of leisure engagement (none, weekly, monthly) to provide a more comprehensive picture of the associations with experiences of aging. Finally, due to concerns around reverse causation, we limited the sample to participants without chronic health conditions at baseline.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Source link

Related Articles

Leave a Reply

[td_block_social_counter facebook="beingmedicos1" twitter="being_medicos" youtube="beingmedicosgroup" style="style8 td-social-boxed td-social-font-icons" tdc_css="eyJhbGwiOnsibWFyZ2luLWJvdHRvbSI6IjM4IiwiZGlzcGxheSI6IiJ9LCJwb3J0cmFpdCI6eyJtYXJnaW4tYm90dG9tIjoiMzAiLCJkaXNwbGF5IjoiIn0sInBvcnRyYWl0X21heF93aWR0aCI6MTAxOCwicG9ydHJhaXRfbWluX3dpZHRoIjo3Njh9" custom_title="Stay Connected" block_template_id="td_block_template_8" f_header_font_family="712" f_header_font_transform="uppercase" f_header_font_weight="500" f_header_font_size="17" border_color="#dd3333"]
- Advertisement -spot_img

Latest Articles