The research, conducted by Johanna Edvinsson, Svend Erik Mathiassen, and David M. Hallman at the University of Gävle in Sweden, addressed a growing paradox in modern knowledge work. Flexible work arrangements, which allow employees to decide when and where they work, are widely promoted as a route to better work-life balance. Yet a substantial body of evidence indicates that flexibility cuts both ways. When the boundary between office and home dissolves, so does the psychological signal that work is finished. An always-on culture can emerge within workgroups, in which employees feel obligated to answer messages late into the evening, extend their working days, and sacrifice both physical activity and sleep. Previous studies have linked this blurring of boundaries to exhaustion, poorer sleep quality, and an elevated need for recovery.

Recovery, in the occupational health literature, refers to the unwinding processes through which strain accumulated during work returns to its baseline level. Psychological detachment, the subjective experience of switching off and leaving work behind, is considered central to this process, particularly in mentally demanding desk jobs. Meta-analyses have associated stronger detachment with less exhaustion, better well-being, and better sleep. But most intervention research in this field has relied on self-reported outcomes, and few studies have asked whether workplace programs can change objectively measured, around-the-clock behavior patterns or the physiology of recovery during sleep. The Swedish team set out to fill that gap with hard measurement data rather than questionnaires alone.

The intervention itself emerged from a co-creation process involving employees, employer representatives, and researchers within a large Swedish governmental agency. Rather than imposing a top-down wellness program, the researchers drew on a comprehensive questionnaire distributed to roughly 4,900 employees and on focus group interviews across all divisions of the organization to identify what workers actually needed. Two activities resulted. The first was an individual-level course, delivered to small groups, that trained participants in using digital tools to plan and prioritize work, minimize interruptions, and end the day with a sense of control. The second was a workgroup-level workshop in which colleagues and managers jointly developed common rules and routines for flexible work, such as how to signal availability to colleagues and what to expect about responding to work communication outside regular hours. Approximately 94 percent of eligible employees took the course and 82 percent attended the workshops, and about a third of workshop participants reported continuing to work on their action plans afterward.

The study used a quasi-experimental design with two departments in the same organizational division, one receiving the intervention and one serving as a control. Twenty-seven intervention participants and 21 controls completed baseline measurements, and 42 of them were followed up twelve months later. Each participant wore a thigh-mounted ActiGraph accelerometer for three consecutive days while also wearing a Firstbeat heart rate monitor that recorded beat-to-beat heart intervals. Sleep periods were identified from diaries, and waking time was classified into active behaviors such as walking, cycling, and stair climbing, versus inactive behaviors such as sitting, lying, and standing. Because the 24-hour day is a fixed budget in which every behavior competes with every other, the team analyzed the data using compositional data analysis, computing isometric log-transformed ratios that capture sleep relative to time awake and, within waking hours, activity relative to inactivity.

The headline result concerned sleep. In the intervention group, nightly sleep rose from an average of 376 minutes at baseline to 412 minutes at follow-up, an increase of 36 minutes that lifted sleep from 26.1 percent to 28.6 percent of the 24-hour day. The control group moved in the opposite direction, dropping from 416 to 393 minutes. Statistical testing confirmed that the difference in change between groups was robust, with a large effect size for the sleep-versus-wake ratio that grew even stronger after adjusting for age, sex, body mass index, and whether participants had children at home. The authors note that the magnitude of the gain is striking when compared against clinical benchmarks: a meta-analysis of randomized trials found that melatonin, a standard treatment for primary sleep disorders, produces sleep-duration improvements of only 5 to 25 minutes. The intervention group’s new average of nearly seven hours per night brought participants meaningfully closer to the seven to nine hours recommended for adults.

Not every outcome moved in the same direction. The intervention did not significantly change the balance of physical activity relative to inactivity during waking hours, and it did not improve heart rate variability during sleep, the study’s physiological marker of recovery. Heart rate variability reflects the interplay between the sympathetic and parasympathetic branches of the autonomic nervous system, with higher values during sleep, particularly the RMSSD and SDNN metrics, indicating greater parasympathetic activity and more efficient physiological restoration. The researchers offer several possible explanations. The intervention was designed to change work practices rather than autonomic regulation directly; physiological adaptations may simply require more time than twelve months of follow-up captured; and the small sample, combined with considerable day-to-day variability in heart rate variability between individuals, may have masked a small but genuine effect. The extra sleep also appears to have displaced both active and inactive waking time in roughly equal proportion, leaving the activity-to-inactivity ratio essentially unchanged.

The findings dovetail with a companion publication from the same project showing that the intervention improved self-reported psychological detachment and reduced the need for recovery, particularly among employees who began the study with the highest recovery needs. Taken together, the two studies suggest that the program successfully supported psychological and behavioral dimensions of recovery, with longer sleep as its clearest objective signature, but that these changes were not sufficient to shift physiological recovery as indexed by cardiac autonomic function. The authors suggest that future interventions aiming at physiological outcomes might incorporate explicit relaxation components, such as mindfulness training, slow breathing exercises, or biofeedback, which previous research has linked to improvements in heart rate variability, alongside the organizational and individual strategies tested here.

The study has limitations that the authors acknowledge candidly. Participants were not randomized to groups, because the intervention had to be scheduled around organizational constraints, and the two departments, although geographically separated, belonged to the same division, leaving open the possibility of contamination through informal conversation. The sample was small, sleep periods were identified from self-reported diaries rather than objective sleep staging, and three days of measurement, while sufficient for reliable estimates of most physical behaviors, cannot capture every nuance of sleep quality, nighttime awakenings, or daytime napping. The co-creation process that made the intervention relevant and feasible in this particular agency may also limit generalizability to other occupational settings.

Even so, the implications are considerable for the millions of knowledge workers now operating under flexible or hybrid arrangements. The results indicate that interventions targeting both individual work practices and collective workgroup norms, developed participatively with the people who must live with them, can produce measurable, clinically meaningful gains in sleep duration over a full year. As employers worldwide grapple with the health consequences of digitalized, boundary-free work, this Swedish experiment offers a concrete template: teach workers to manage their tools and priorities, and let teams negotiate shared rules about availability, and the 24-hour distribution of rest may rebalance itself toward recovery. Longer-term studies with larger samples, additional biomarkers such as cortisol and blood pressure, and richer sleep assessment will be needed to determine whether the behavioral gains translate into physiological and health benefits, but the demonstration that workplace design can add half an hour of sleep per night is, by itself, a result worth losing sleep over.

Subject of Research: Effects of a co-created workplace intervention on sleep, 24-hour physical behaviors, and heart rate variability in office workers with flexible work arrangements.

Article Title: Effects on 24-h physical behaviors and heart rate variability during sleep of a co-created workplace intervention to promote recovery in office workers with flexible work

Article References: Edvinsson, J., Mathiassen, S. E., & Hallman, D. M. (2026). Effects on 24-h physical behaviors and heart rate variability during sleep of a co-created workplace intervention to promote recovery in office workers with flexible work. Journal of Activity, Sedentary and Sleep Behaviors, 5(1), Article 14. https://doi.org/10.1186/s44167-026-00107-0

Image Credits: AI Generated

DOI: 10.1186/s44167-026-00107-0

Keywords: flexible work, workplace intervention, sleep, recovery, heart rate variability, physical activity, sedentary behavior, office workers, occupational health, co-creation, compositional data analysis, psychological detachment