A longitudinal cohort study tracking older Swedish adults across their transition away from full-time employment has revealed that retiring reshapes the overall composition of daily movement behaviors, with pre-retirement workplace physical demands playing a decisive role in post-retirement activity patterns. Drawing upon device-based measurements collected before and after labor market exit, researchers observed that the shift into retirement led to an overall increase in sleep duration relative to waking hours across the entire sample. However, the subsequent shifts in sedentary behavior and physical activity were strongly dictated by the amount of physical activity participants had performed on the job, resulting in an equalization that diminished pre-existing behavioral disparities between occupational groups.
Published in the European Review of Aging and Physical Activity, the investigation evaluated data from the Swedish Retirement Study to clarify how older adults redistribute their time across the complete twenty-four-hour cycle. While previous research has linked retirement with shifts in physical activity, prior investigations frequently relied on self-reported measures or examined behaviors in isolation, yielding contradictory results regarding whether retirement primarily drives healthier active routines or increases sitting time. To address these gaps, the study authors utilized compositional data analysis, an analytical framework specifically designed to evaluate mutually exclusive and co-dependent fractions of a finite twenty-four-hour day, ensuring that changes in one movement state are mathematically contextualized against reciprocal shifts in others.
The study tracked 112 participants, comprising 47 men and 65 women aged between 60 and 72 years, across three distinct assessment timepoints: approximately six months prior to retirement, and at six months and eighteen months following retirement. Participants were recruited from earlier survey waves of the Swedish Longitudinal Occupational Survey of Health and the Aging at Work study. To qualify for inclusion, participants were required to be working full-time before transitioning, reduce their working hours by at least 50 percent, and subsequently work no more than 25 percent of a full-time schedule. Individuals who engaged in regular nightshift work between 10:00 pm and 6:00 am, those taking prolonged sick leave, or those facing unemployment in the months immediately preceding retirement were excluded from the analysis.
To obtain reliable, objective records of movement behaviors, the research team deployed a combination of body-worn devices alongside detailed participant diaries over week-long collection periods at each timepoint. Sleep onset and final morning awakening were derived using wrist-worn actigraphs placed on the non-dominant arm and processed via the Cole-Kripke algorithm. Meanwhile, daily waking postures and movements were measured using triaxial accelerometers affixed to the thigh. Using the ActiPASS automated software, raw acceleration signals were classified into distinct postural and movement states according to established guidelines, separating daily life into sleep, sedentary behavior, light physical activity, and moderate-to-vigorous physical activity.
Pre-retirement occupational physical activity was quantified by cross-referencing daily diary entries of work shift start and end times against thigh accelerometer records. Participants were stratified into tertiles based on the percentage of their working hours spent moving in light, moderate, or vigorous activity. The low occupational physical activity group spent an average of 18.7 percent of their working time active, typified by desk-based and sedentary occupations. The medium tertile averaged 36 percent active work time, while the high occupational physical activity group spent an average of 63 percent of work hours engaged in physically demanding tasks such as standing, walking, carrying, and manual labor.
Baseline characteristics indicated an average participant age of 64.9 years, with 81 percent of the cohort married or cohabiting and 71 percent having worked in managerial roles or occupations requiring higher education. In addition, 61 percent of participants reported managing one or more physician-diagnosed chronic health conditions that affected their daily life. Descriptive compositional summaries showed that across all measurement occasions, sedentary behavior accounted for the largest single share of the 24-hour day, averaging 10.7 hours daily before retirement, 10.35 hours daily at six months post-retirement, and 10.55 hours daily at eighteen months post-retirement, followed by sleep, light activity, and moderate-to-vigorous exertion.
By transforming raw time allocations into unconstrained isometric log-ratio coordinates, the authors fitted multivariable linear mixed models adjusted for sociodemographic factors, baseline chronic illness, and occupational class. Across the cohort as a whole, retirement produced a statistically significant increase in the ratio of sleep to total wake time, observed at both the six-month follow-up (beta: 0.08, 95% CI: 0.04 to 0.12) and the eighteen-month follow-up (beta: 0.07, 95% CI: 0.03 to 0.11). When viewing the entire cohort without occupational stratification, no statistically significant overall changes were detected in the balance between sedentary time and total physical activity, or in the ratio between light and moderate-to-vigorous physical activity.
Marked divergence emerged, however, when the linear mixed models incorporated interaction terms between retirement status and pre-retirement occupational activity levels. The rise in sleep relative to wake time was driven predominantly by retirees from the high occupational activity group, who experienced substantial gains in sleep relative to their waking day. To a lesser degree, participants in the medium occupational activity group also demonstrated an increase in sleep relative to waking time at six months post-retirement, whereas sleep ratios among workers retiring from sedentary jobs remained comparatively stable across the transition.
The most striking behavioral reorganization occurred in the allocation between sedentary behavior and total physical activity, where the high and low occupational groups exhibited diametrically opposing shifts that led to behavioral convergence. Workers exiting physically demanding occupations significantly increased their ratio of sedentary time relative to physically active time following retirement. Conversely, individuals transitioning out of desk-bound, sedentary jobs significantly decreased their sedentary-to-active ratio, moving more during their newly acquired post-retirement leisure hours and closing the substantial pre-retirement activity gap that previously separated the two groups.
The study authors interpreted these behavioral shifts through the frameworks of the physical activity paradox and the sweet-spot hypothesis of 24-hour movement balance. While occupational physical activity frequently entails static posture, prolonged standing, heavy lifting, and cardiovascular strain without adequate recovery, leisure-time physical activity typically offers restorative aerobic training effects. Consequently, the observed reduction in overall physical activity and corresponding increase in sedentary time and sleep among formerly heavy-labor workers does not necessarily signify an unhealthy decline in functional vitality, but rather a restorative rebalancing away from excessive workplace physical strain toward their personal physiological sweet spot.
The researchers also identified key methodological and demographic limitations within their investigation. The analytic sample comprised 112 participants drawn exclusively from Sweden, which may limit generalizability to other labor markets or countries with distinct retirement systems and cultural retirement norms. Furthermore, participants who opted into the cohort study were planning their retirement at normative ages and were relatively healthy and functional, leaving open questions regarding whether identical behavioral trajectories occur among individuals who exit work prematurely due to sudden disability, involuntary redundancy, or chronic health complications.
Additional measurement constraints arose from the data collection parameters, which categorized daily life into four broad movement behaviors and excluded daytime naps from the sleep construct because actigraphy was calibrated exclusively to nighttime sleep intervals. Commuting time to and from workplaces was also unmeasured, which could influence total daily energy expenditure estimates prior to labor market exit. The authors emphasized that future research employing larger, socioeconomically diverse cohorts will be necessary to examine varied pathways into retirement and determine how post-retirement activity reallocations influence long-term cardiometabolic and musculoskeletal health outcomes.
Subject of Research: Changes in 24-hour movement behaviors across retirement evaluated through compositional data analysis
Article Title: Changes in physical activity across retirement: a compositional data analysis approach in a Swedish cohort study
Article References: Sacco, L. B., Högnäs, R. S., Palarea-Albaladejo, J., Hettiarachchi, P., Svartengren, M., & Westerlund, H. (2025). Changes in physical activity across retirement: a compositional data analysis approach in a Swedish cohort study. European Review of Aging and Physical Activity, 23(1), Article 3. https://doi.org/10.1186/s11556-025-00395-6
Image Credits: AI Generated
DOI: 10.1186/s11556-025-00395-6
Keywords: Retirement transition, Movement behaviors, Compositional data analysis, Occupational physical activity, Sedentary behavior, Sleep duration, Light physical activity, Moderate-to-vigorous physical activity, Swedish Retirement Study, Thigh accelerometry, Wrist actigraphy, Physical activity paradox, Sweet-Spot Hypothesis, Time-use epidemiology, Linear mixed models, Aging and health, Isometric log-ratio
News Source: Ophelia Keating. (November 13, 2026). Retirement Equalizes Daily Movement Patterns Across Different Occupational Activity Levels. Scienmag.
