Teleworking and its impact on health and well-being: a systematic review of longitudinal studies considering the psychosocial work environment (2005–2024)

Abstract

Background

Teleworking expanded dramatically during the COVID-19 pandemic and is expected to remain widespread. While it offers benefits such as greater autonomy and improved work–life balance, teleworking also poses challenges related to workload, social isolation, and blurred boundaries between work and personal life, shaped by the psychosocial work environment. These contrasting effects raise questions about whether teleworking promotes or undermines workers’ health. This systematic review synthesizes longitudinal evidence on the relationship between teleworking and workers’ health, with particular attention to the role of the psychosocial work environment as an explanatory framework. Specifically, it addresses two questions: (1) What is the impact of teleworking on health (physical and mental) when psychosocial work environment factors are considered? (2) How have longitudinal studies examined the relationship between teleworking and health, including direct, mediation, and moderation effects, and how the role of teleworking in these relationships has been studied?

Methods

This systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42024599518). We searched five electronic databases (Embase, Medline, PsycInfo, Web of Science, CINAHL Plus) and bibliographies for longitudinal studies published between 2005 and 2024 in OECD countries. Eligible studies examined teleworking and workers’ health (physical and/or psychological) in combination with psychosocial work environment factors. Study quality was assessed using the National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment Tool. Data from 20 studies published between 2017 and 2024 were extracted and synthesized based on study and psychosocial work environment characteristics, as well as the nature of the association between telework and health outcomes.

Results

All 20 studies were rated fair or good in methodological quality. Teleworking effects on health were heterogeneous and context-dependent, increasing psychological distress, exhaustion, and anxiety under high psychosocial risk (i.e., adverse work conditions such as high workload or work intensification, low autonomy, limited social support, job insecurity, reduced psychological detachment, and blurred work–nonwork boundaries) or mandatory arrangements, but protective with autonomy, supportive psychosocial conditions, or individual resources. Psychosocial work environment factors such as autonomy, job security, workload, and social support emerged as central determinants of health outcomes. Effects operated through moderation and mediation involving psychosocial work environment factors. Important gaps remain, particularly the limited investigation of physical health outcomes.

Conclusions

Teleworking is neither uniformly beneficial nor harmful. Its health impact depends on psychosocial work environment factors, individual characteristics, and contextual conditions. Longitudinal research should expand the range of psychosocial work environment factors examined, consider physical health outcomes, and investigate effects across pre-, during-, and post-pandemic periods. These insights are critical to inform evidence-based workplace interventions and organizational policies that promote employee health.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12889-026-28220-4.

Background

Teleworking, defined as performing work activities outside the traditional office environment while relying on digital or telecommunication technologies to stay connected to the organization [1], is used in the present review as a broad and inclusive term encompassing working from home, hybrid work arrangements, and other forms of remote work. Although teleworking during the coronavirus disease (COVID-19) pandemic most often referred to working from home arrangements, the broader term “teleworking” was retained to reflect the diversity of terminology and work arrangements examined across studies published over time.

Teleworking is not a new phenomenon. First introduced in the 1970 s as an alternative form of work organization facilitated by advances in information and communication technologies, it remained relatively marginal for several decades, mainly limited to specific occupations and industries [2]. However, teleworking expanded on an unprecedented scale with the onset of the COVID-19 pandemic, largely due to public health measures aimed at limiting physical interactions [3, 4]. This trend has been observed globally: in the US, the share of paid work-from-home days increased from 5–7% before the pandemic to around 28% by mid-2023 [5], while in Europe, occasional or regular telework doubled from 11% pre-pandemic to 22% in 2021 [6]. Similarly, in Canada, more than 30% of workers were teleworking at the beginning of 2021, compared to only 4% before the crisis [7]. Despite highly mediatized return-to-office mandates and narratives announcing the end of telework, global data indicate a stabilization of telework arrangements since 2023 [5]. Available data suggest that both hybrid telework and intensive telework (≥ 3 days per week) are likely to persist in the longer term [7, 8] and could become the norm for 20–25% of workers worldwide [9]. More recently, empirical evidence highlights the increasing implementation and complexity of return-to-office (RTO) policies across organizations [10]. For example, the Canadian federal government has announced a phased return to full in-office work by 2027 [11], while major technology firms such as Google have required some remote employees to transition to hybrid schedules [12]. These developments indicate that the long-term trajectory of telework remains dynamic and context-dependent, and that telework prevalence may continue to fluctuate across sectors and jurisdictions.

Accompanying the rapid growth of remote work, research has also expanded to explore its consequences for organizations and workers. For organizations, telework generates a dual set of effects. On the one hand, it can enhance organizational performance by increasing employees’ self-reported productivity, particularly when tasks are independent, workers enjoy autonomy and telework is voluntary rather than imposed [13]. Studies also show that telework may yield measurable gains in output and efficiency for roles involving standardized tasks, while increasing flexibility, reducing operational costs related to office space and physical infrastructure, and potentially improving environmental performance [13, 14]. On the other hand, telework can create challenges in managing remote employees, require substantial reorganization of business processes, and lead to coordination difficulties or weaker collaborative networks. It may also generate higher organizational costs associated with technology provisioning, cybersecurity, telecommunication equipment, and employee training [14]. For workers, teleworking offers numerous advantages, promoting work-life balance, alleviating perceived stress, enhancing autonomy, and fostering greater concentration and flexibility at work [15–18]. However, it also entails several drawbacks, including blurred boundaries between professional and personal life [17–19], difficulties in social and professional integration [18], heavier workloads and longer working hours, and the perceived need to remain constantly available to maintain a positive image [20]. In addition, teleworking may reduce workers’ access to social support in the workplace [18] and increase feelings of isolation [15]. Similarly, difficulty in obtaining support from supervisors can have an impact on engagement and productivity and can hinder knowledge sharing [21]. From a social interaction perspective, however, working on-site also involves certain costs. Researchers have highlighted the issue of collaboration overload [22]. While social interaction can reduce autonomy [23], it accounts for nearly 60% of work interruptions [24], which require effortful task switching, self-regulation, and emotional labor. These processes heighten cognitive arousal, deplete energy, and may contribute to work exhaustion [25]. Moreover, teleworkers may be perceived negatively in relation to their commitment and contribution by their non-teleworkers colleagues, generating an “us/them” dynamic of exclusion [26].

While these findings highlight both benefits and drawbacks of teleworking for workers, its overall effects remain difficult to capture, particularly regarding workers’ health. Mixed findings may reflect a lack of consideration of the broader work environment simultaneously affecting health. Recent systematic reviews have similarly emphasized this heterogeneity of findings (e.g., [27, 28]), suggesting that the relationship between teleworking and health is complex and likely dependent on contextual work conditions rather than telework per se.

In this context, the concept of the psychosocial work environment is particularly relevant. Drawing on the World Health Organization framework (e.g., [29]) and comprehensive syntheses of psychosocial work exposures (e.g., [30]), it encompasses the organizational, social, and psychological conditions under which work is performed. This concept covers a wide range of factors, including job demands, job control, social support, recognition, job insecurity, workplace bullying, work–family conflict, organizational injustice, and leadership quality. These factors reflect key dimensions of work design, organizational context, and interpersonal relationships that influence workers’ physical and mental health. This framework is embedded within broader Occupational Health and Safety (OHS) approaches that emphasize the prevention and management of work-related risks, including psychosocial risks, as core determinants of workers’ health.

This perspective is consistent with major theoretical models in occupational health research, including the Job Demand-Control model [31], the Effort-Reward Imbalance model [32], and the Job Demands-Resources model [33], which emphasize the joint role of job demands and resources in shaping health outcomes. In this systematic review, the psychosocial work environment is used as an overarching analytical framework guiding study selection, data extraction, and synthesis. Only studies explicitly integrating teleworking with psychosocial work environment factors were included, and variables were systematically categorized into job demands, job resources, and interpersonal and/or organizational factors.

However, important methodological limitations remain in the existing literature. Cross-sectional studies, which constitute most of the existing research, limit the ability to establish causal relationships and to capture how the effects of teleworking on health evolve over time. Despite increasing research, especially since the COVID-19 pandemic, important gaps remain. Existing systematic reviews (e.g., [27, 28]) have not simultaneously focused on teleworking and the psychosocial work environment as an integrated explanatory framework, nor on longitudinal mechanisms such as direct, moderating, or mediating effects. The present systematic review addresses these shortcomings.

A clearer understanding of these relationships is crucial for informing organizational policies, workplace interventions, and public health strategies aimed at promoting workers’ health. Within this conceptual context, this systematic review synthesizes longitudinal evidence on the relationship between teleworking and health, explicitly considering the psychosocial work environment as an organizing analytical framework. Accordingly, it aims to address two main research questions: (1) What is the impact of teleworking on health (physical and mental) when the psychosocial work environment is considered? and (2) How have longitudinal studies examined the relationship between teleworking and health, including direct, mediation, and moderation effects, and how the role of teleworking in these relationships is addressed?

Regarding physical health consequences, unplanned teleworking during the COVID-19 pandemic has been associated with increased musculoskeletal disorders (MSDs) due to the lack of ergonomic equipment and dedicated workspace [14]. Another study highlighted that teleworkers’ posture was linked to the occurrence of MSDs, with the shoulders, neck, and wrists being the most affected regions [34]. To the best of our knowledge, no further studies have investigated this issue, highlighting the limited research on the physical health consequences of teleworking.

With respect to mental health, findings reveal mixed effects. Evidence from a recent systematic review indicates that teleworking has a dualistic nature [35]. On the one hand, several studies report favorable experiences associated with remote work. On the other hand, other studies included in this systematic review do not support the idea that this work arrangement leads to sustained positive outcomes. A meta-analysis of observational studies further highlighted the main adverse effects of teleworking [36]. Although most studies included in this meta-analysis reported a deterioration in workers’ mental health, some also highlighted potential benefits. These findings underscore the need for further research to clarify the causal mechanisms underlying the relationship between teleworking and mental health outcomes.

Taken together, these considerations raise an important question: Do the benefits of teleworking truly outweigh its drawbacks? If not, teleworking may evolve into an emerging occupational risk factor, amplifying both physical and mental health problems.

The expansion of teleworking as a major shift in work organization has itself been associated with increased psychosocial risks (PSR), which constitute adverse manifestations of the broader psychosocial work environment. Exposure to high workload, low autonomy, poor social support, or limited recognition are well-established determinants of mental and physical health at work [37]. While teleworking could mitigate some risks, it may exacerbate others. Employees who worked from home during the pandemic generally reported more positive psychosocial work environment factors and better health outcomes, whereas those facing changes in work tasks or team reorganizations experienced more negative effects [38].

Furthermore, evidence suggests that changes in composite indicators of job stress are mainly driven by variations in psychosocial stressors rather than resources, highlighting that modifications in work organization may primarily operate through changes in exposure to job demands and constraints [39]. According to the authors, these findings suggest that working from home may have entailed fewer disturbances and distractions, thereby suggesting that changes in psychosocial stressors may play a central role in shaping health outcomes in teleworking contexts.

Methods

This systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42024599518) [40] and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol [41].

Definition and eligibility criteria

This systematic review focused on quantitative longitudinal studies examining the relationship between teleworking and workers’ physical or mental health. For the purposes of this review, teleworking was defined as the performance of job tasks outside the employer’s premises, regardless of the technology used for at least one day per week, whether on a part-time or full-time basis [1]. Telework activities could take place in any setting, in any professional category, in any segment of the labor market.

The inclusion criteria were as follows: (a) Population: the review targeted employees aged 18 years and older who were employed in the Organisation for Economic Co-operation and Development (OECD) member countries. This restriction was applied to ensure that the populations under study shared comparable economic, regulatory, and occupational health contexts, thereby minimizing variability related to informal labor markets or differing welfare systems; (b) Exposure: studies that assessed any form of teleworking, as previously defined, were included. Furthermore, only studies that examined telework in combination with psychosocial work environment factors, conceptually defined as organizational, social, and psychological working conditions such as job demands, job control, social support, and recognition in line with the World Health Organization framework and syntheses of psychosocial exposures (e.g., [29, 30]), were included to ensure alignment with the research question; (c) Outcomes: studies that stated the outcome variable was mental health and/or physical health including well-being regardless of how it is measured; (d) Study characteristics: eligible studies were limited to quantitative longitudinal observational designs published in peer-reviewed journals in English between January 1, 2005, and December 31, 2024. This time frame was selected to coincide with the widespread diffusion of broadband and digital technologies that have made telework a viable and sustained mode of employment in OECD economies [42], as well as to capture the acceleration and institutionalization of teleworking practices during and since the COVID-19 pandemic.

Search strategy

Five electronic databases were queried: Embase (Ovid), Medline (Ovid), PsycInfo (Ovid), Web of Science (Clarivate), and CINAHL Plus (EBSCO). Searches covered the period from January 1, 2005 to December 31, 2024. Search strategies combined controlled vocabulary with free-text keywords, capturing three main concepts: (a) teleworking, (b) health outcomes and (c) study design. The search strategy was developed within a psychosocial work environment framework, which guided the conceptualization of the review, study selection, inclusion criteria, and interpretation of findings. Database-specific syntax adjustments were applied. The complete search strategies are documented in Additional file 1. In addition, a confirmatory search strategy was conducted based on an inductive screening of bibliographies from included studies (backward citation searching), as well as articles citing the later (forward citation searching, i.e., snowballing).

Data extraction and management

Data extraction and management followed a structured, multi-stage approach using Covidence, a web-based platform that streamlines systematic reviews through workflows [39, 43]. Once all electronic bibliographic databases were queried, a total of 6,009 references were imported into Covidence. After removing 1,538 duplicates, 4,471 titles and abstracts were independently screened by two reviewers (MEB, OR) for eligibility. Strong inter-rater agreement was achieved (proportionate agreement = 0.92, random agreement of 0.84). As part of the full-text screening phase, the exposure definition was refined to ensure conceptual precision, limiting inclusion to studies that examined teleworking in combination with psychosocial work environment variables within the same analytical model. The research team (NB, OR, MEB, APL) discussed potentially eligible studies that did not clearly meet these criteria to resolve uncertainties and confirm eligibility. Following this process, 20 studies were retained and constituted the final analytical sample for the systematic review.

Quality assessment

The National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment Tool designed for observational cohort and cross-sectional studies was used to assess the quality of the included studies [44]. Two reviewers (MEB, NB) independently evaluated the 14-item tool for each included study, any discrepancies in the assessments were discussed and solved by a third reviewer (APL). The scoring system reinforces the presence of key components in longitudinal designs (e.g., loss to follow-up) and clarity in methodological reporting (Yes = 1; Not = 0.5; Not reported/Cannot determine/Not applicable = 0). Studies were classified as Poor (score ranging from 0–7), Fair (score ranging from 7.5–10.5), or Good (score ranging from 11–14). Missing data was dealt with reviewing other publications from the same studies, no authors were contacted accordingly. Data was further synthesized based on study and psychosocial work environment characteristics, as well as the nature of the association between telework and health outcomes in alignment with the research questions.

Results

Study characteristics

We identified 6,009 studies from the original literature search. Out of the 337 studies initially considered potentially relevant, 18 were retained for review [45–62]. Exclusion criteria included inappropriate study setting, outcomes, exposure, design, or publication type. An additional eight studies were identified through the confirmatory search (i.e., backward and forward citation searching, or snowballing) (full details can be obtained from the authors upon request). Of these, six were excluded because of unsuitable outcomes, exposures, or study design. The remaining two were included [63, 64], bringing the final number of studies reviewed to 20. The screening and selection procedure is illustrated in Fig. 1.

Fig. 1.

PRISMA 2020 flow diagram of the study selection process. Flow diagram illustrating the identification, screening, eligibility assessment, and inclusion of studies retrieved from database searches and additional sources. *Consider, if feasible to do so, reporting the number of records identified from each database or register searched (rather than the total number across all databases/registers). **If automation tools were used, indicate how many records were excluded by a human and how many were excluded by automation tools. Source: Page MJ, et al. BMJ 2021;372:n71. https://doi.org/10.1136/bmj.n71. This work is licensed under CC BY 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/

The final set of studies, published between 2017 and 2024, comprised 20 longitudinal designs with different follow-up structures, most involving two measurement points and time lags ranging from one week to 84 months. No eligible longitudinal studies meeting the inclusion criteria were identified prior to 2017 despite the search covering the period from 2005 onward. Table 1 summarizes their main characteristics. All included studies were conducted in OECD countries, the majority in Europe [45–54, 56, 58–60, 63, 64], along with two from the United States [55, 61], one from Japan [57], and one from Chile [62]. Each study controlled for sociodemographic characteristics, most often gender and/or age, while fewer accounted for additional factors such as personality traits, health habits (e.g., alcohol consumption), family status (e.g., parental status), and other work factors (e.g., prior telework experience). Almost all studies focused on mental health outcomes, with four exceptions: one that examined self-rated health, defined as the individual’s overall perception of health [64]; two others investigated psychosomatic symptoms, described as fatigue [47], as well as psychological and physical stress responses [57]; and a fourth that assessed work ability [50]. The latter was described by the authors as an individual’s capacity to perform work tasks physically, mentally, and socially, encompassing physical health, psychological resilience, vocational skills, and the ability to meet job demands [50]. None of the studies specifically investigated physical health outcomes. A wide range of indicators was used to measure mental health. The most frequently assessed construct was exhaustion (n = 10), followed by psychological distress (n = 2), COVID-19–related anxiety (n = 2), depressive symptoms (n = 2), while fewer studies examined general well-being (n = 1), general anxiety (n = 1), or broader mental health (n = 1). Some studies assessed more than one mental health outcome simultaneously.

Table 1.

Summary of the 20 quantitative studies

Author (Ref. number)	Country	Sample characteristics	Sample (N)		Health measure	Findings				Quality	Type of effect	Telework impact
						Teleworking	Psychosocial work environment factors	Other factors	Covariates
De Carlo et al., 2022 [46]	Italy	Workers in general	185		Exhaustion (Maslach Burnout Inventory—Emotional Exhaustion scale, 9 items)	Work arrangement: ns/Workload x Work arrangement: +/Social support x Work arrangement: +	Workload: ns/Social support: −	T1 outcome: +	Gender, age, contract type	Fair	Moderating effect of teleworking on the relationship between workload and exhaustion and on the relationship between social support and exhaustion	Generally unfavorable effect of teleworking
Perry et al., 2018 [55]	United States	Full-time working adults (Study 1)	258	Exhaustion (Oldenburg Burnout Inventory)		Remote work (A): ns/A × B: ns/A x C: ns/A × B × C: ns	Autonomy (B):—/B × C: -	Emotional stability (C): -	Gender, hours worked/week, salary, organizational tenure, permanency of work arrangement	Fair	Direct effect. Moderating effect of teleworking: a high level of remote work causes exhaustion, but the effect is moderated by a high level of autonomy	Generally non-significant effect of teleworking, unless autonomy is low, in which case teleworking is unfavorable
Shimura et al., 2021 [57]	Japan	Office workers from tertiary industries	3123	Psychological and physical stress response/Psychosomatic symptoms (Brief Job Stress Questionnaire – BJSQ)		1–2 days/week:—/3–4 days/week:—/5 days/week (“full-remote”): -	Job stressors: +/Social support (colleagues & supervisors): -	Sleep disturbance: +/Total sleep time on workdays (hours): +/Total sleep time on free days (hours): ns	Age, gender, overtime work	Fair	Direct effect	Generally favorable effect of teleworking
Van Zoonen & Sivunen, 2022 [59]	Finland	Finnish workers mostly from the public administration 45.3%; 4.6% manufacturing	1164	Psychological distress (5 items from GHQ-12)		Remote work → isolation + → Distress +	Isolation T1 → Distress T2: +/Distress T1 → Distress T2: +/Isolation T2 → Distress T3: +/Distress T2 → Distress T3: +/Mediated Communication T1 → Isolation T2:—/Isolation T2 → Distress T3: -	n/a	Household size, remote office characteristics, age, gender, and organizational tenure, remote work before the pandemic	Fair	Indirect/Mediation effect of teleworking	Generally unfavorable effect of teleworking
Windeler et al., 2017 [61]	United States	Workers from a financial services firm	51	Exhaustion (Maslach Burnout Inventory—GS)			Before part-time telework: Interpersonal interaction + Job interdependence + External interaction ns After part-time telework: Interpersonal interaction ns Job interdependence + External interaction +	n/a	Gender, age	Fair	Moderating effect of part-time teleworking. PTT weakens the positive relationship between interpersonal interaction and work exhaustion, but strengthens the positive relationship between external interaction and work exhaustion. PTT provides recovery from interpersonal interaction but not external interaction; after PTT, work exhaustion increases as external interaction increases	Generally favorable effect of teleworking
Meyer et al., 2021 [53]	Germany	Workers in general	789	Exhaustion: Copenhague Burnout Inventory		Home office (0 = yes, 1 = no):—/Home office (0 = yes, 1 = other): ns/Home office (0 = yes, 1 = both from home and on site): -	Co-worker social support:—/Job insecurity: +	n/a	Age, gender	Fair	Direct effect	Unfavorable effect when teleworking is experienced as forced, as during the pandemic
Savolainen et al., 2021 [56]	Finland	Workers participating to the Social Media at Work in Finland Survey (SMWFS)	1044	Covid Anxiety (scale adapted) scale of the Spielberger State–Trait Anxiety Inventory STAI-6		Remote work: ns	Technostress: +/social support from work: ns	Work exhaustion: ns/Female: +/Age: ns/High income: ns/University degree: ns/Occupational area: ns/Psychological loneliness: +/Psychological distress: +/Openness: ns/Conscientiousness: ns/Extroversion: ns/Agreeableness: ns/Neuroticism: +/Social media information bubble: ns/	n/a	Fair	Direct effect	Generally unfavorable or non-significant effect of teleworking
Vanderstukken et al., 2022 [60]	Belgium	Workers from 9 organizations	564	Emotionnal Exhaustion: UBOS (5 items)		Job requirement profile: ns/Efficiency profile:—/Work-life balance profile: -	Speed of work: ns/Lack of clarity: ns/Task problems: ns/Task variety: ns/Autonomy:—/Job security/ns	T1 outcome: ns	Gender, age, level of education	Good	Direct effect of latent classes according to the reasons for teleworking	Favorable effect for two latent classes and insignificant for a third latent class
Celuch et al., 2024 [45]	Finland	Workers participating to the Social Media at Work in Finland Survey (SMWFS)	768	Psychological distress (GHQ-12), Work Exhaustion (MBI)		Remote work: ns	Cyberbullying victim at work: + regarding (Psychological distress) but ns regarding (Work exhaustion)	Social media identity bubble: ns/Compulsive social media use: +	Gender, age, education level, managerial position, personality, occupational area	Good	Direct effect	Non-significant effect of teleworking
Kröner & Müller, 2023 [51]	Germany	Workers in general	642	Well-being: WHO (5 items),		Telework: ns/Telework × Occupational self-efficacy: +/Telework × Social support: ns/Telework × Decision-making autonomy: ns/Telework × Flexitime: ns/Telework × Work intensification:—(only for Medium well-being)	Work intensification: +/Social support: ns/Decision-making autonomy: ns/Flexitime: ns/	Occupational self-efficacy: -	Age, gender, job tenure, children, cohabitants, telework experience	Fair	Direct effect. Moderating effect of teleworking. Positive moderating effect: teleworking mitigates the negative effects of work intensification and compensates for a low level of self-efficacy	Favorable effect of teleworking
Urbanavičiūtė et al., 2023 [58]	Lithuania	White-collar employees from different organisations and sectors	236	Exhaustion: Oldenburg Burnout Inventory (4 items)		Telework days more days:—segregative profile/(1–1) Home-protective: ns/(2–2) Work-protective:—/(3–3) Segregative: +/(4–4) Integrative: ns/(1 → 3) Home → Segregative: +/(2 → 3) Work → Segregative: +	Supervisory role: + home-protective (1–1) & integrative (4/4) profiles/Obligation to respond: + work-protective (2/2) profile	n/a	Gender: ns/Age: + work-protective (2–2) profiles/education: + home-protective (1–1) & work-protective (2–2) profiles & – segregative (3–3)/Sector: + work-protective (2–2) profile/Full-time Contract: + integrative (4–4) & work-protective (2–2) profiles	Fair	Direct links between certain border management profiles and exhaustion, but indirect effects between certain factors (e.g., the obligation to respond) and exhaustion via the border management profile	Indirect positive effect. Employees who telework more often are less likely to adopt a "segregative" profile, which is associated with increased burnout over time. Therefore: more teleworking = less segregation = less burnout (indirectly). This positive effect is contingent on the boundaries being managed flexibly
Lazauskaite-Zabielske et al., 2023 [52]	Lithuania	White-collar employees	236	Exhaustion: Oldenburg Burnout Inventory (4 items)		Overwork climate T1 → Overwork climate T2: +/Detachment T1 → Detachment T2: +/Exhaustion T1 → Exhaustion T2: +/Overwork climate T1 → Detachment T2:—(only for telework full-time)/Detachment T1 → Overwork climate T2: ns (only for Telework part-time)/Detachment T1 → Exhaustion T2: ns (only for Telework part-time)/Overwork climate T1 → Exhaustion T2: ns/Overwork climate T1 → Detachment T1: ns (only for Telework part-time)	n/a	n/a	Age: ns/gender: ns	Fair	Moderating effect of the intensity of teleworking on the relationship between overwork climate, psychological detachment and exhaustion	Indirect adverse effect. The authors studied how the overwork climate influences burnout among teleworkers, via the difficulty of psychological detachment from work
Kausto et al., 2024 [50]	Finland	Public sector employees from 11 cities participating to the Finnish Public Sector Survey (FPS)	51 746	Work ability score (WAS)		Working from home (less) Suboptimal decreasing over time trajectory):—/Working from home (more)Relatively good consistent trajectory: +/Consistent optimal trajectory (ref.)	Work time control:—/job strain: +/Team climate:—but ns for Consistent optimal work ability/Organisational justice:—/New tasks: ns/Reorganisation: +	n/a	Age, sex, socioeconomic status (SES), job tenure, employment type, Body Mass Index (BMI), alcohol use, smoking status, physical inactivity	Good	Direct effect of teleworking combines with other work conditions (job strain, team climate, organizational justice, reorganization)	Favorable effect of teleworking with other work conditions
Galliker et al., 2024 [49]	Switzerland	Workers from the LINK Internet Panel	1 553	1) Exhaustion: Oldenburg Burnout Inventory (8 items); 2) Depressive symptoms (CSE 2 items); 3) Irritation; 4) Somatic complaints; 5) Subjective Health; 6) Sleep problems		WfH days before and since COVID-19 (current and additional WfH): 1) 2) 3) 4) 5) 6) ns/Experienced home-office workers → JSI:—(interaction JSIxWfH)	Job stress index (JSI) baseline: + and—for general health & job satisfaction/JSI difference (2021–2020): 1)2)3)4)6) + and 5)- for general health & job satisfaction/Short-term work: 1)3) -; 2)4)5)6) ns	COVID disease: 1)2)3)5) ns 4)6) +/COVID-specific demands index: 1)2)3)4)6) + for almost outcomes and 5) -	Gender, age, language region, education, leadership function, part-time work, children in the household, and home demands	Good	Direct effect of teleworking and gender moderation for health outcome during COVID-19 epidemic	Non-significant effect of teleworking
de Wind et al., 2021 [47]	Netherlands	Workers from the Netherlands Working Conditions Survey (NWCS)	1 168	Fatigue: 5 items-scale		WfH mismatch: ns	n/a	Fatigue at baseline: + (c)	Age: ns; Sex (female): c; Educational level: c	Good	Direct effect	Non-significant effect of teleworking
Dragano et al., 2022 [48]	Germany	Workers from the German National Cohort (NAKO) Study	161 787	Depressive symptoms: PHQ-9; Anxiety: GAD-7		Working from home (yes): PHQ-9: +; GAD-7: +;	Working – change of job: PHQ-9: +; GAD-7: +/Working increased working hours: PHQ-9: +; GAD-7: +/Working reduced hours with short-time allowance: PHQ-9: -; GAD-7:—/Working reduced hours without short-time allowance: PHQ-9: +; GAD-7: +/Job insecurity (high): PHQ-9: +; GAD-7: +	Unemployed prior to Corona: PHQ-9: +; GAD-7: +/Unemployed due to Corona: PHQ-9: +; GAD-7: +/Inactive person: PHQ-9: +; GAD-7: +/Financial situation Improvement: PHQ-9: -; GAD-7:—/Financial situation deterioration: PHQ-9: +; GAD-7: +	Age (19–29 years: ref.), 30–39 years: PHQ-9: +; GAD-7: ns, 40–49 years: PHQ-9: +; GAD-7: +, 50–59 years: PHQ-9: ns; GAD-7: +, 60–69 years: PHQ-9: -; GAD-7: ns, 70–79 years: PHQ-9: -; GAD-7:—/type of household (one person: ref.), multiple-person household without children (< 14 years): PHQ-9: -; GAD-7: +, multiple-person household with children (> 14 years): PHQ-9: -; GAD-7: +/high-risk contact with an infected person: PHQ-9: +; GAD-7: +/tested positive SARS-CoV-2: PHQ-9:-; GAD-7:-/General health (poor): PHQ-9: +; GAD-7: +	Good	Direct effect	Unfavorable effect of teleworking
Oksanen et al., 2022 [54]	Finland	Workers participating to the Social Media at Work in Finland Survey (SMWFS)	656	Covid-19 Anxiety		Remote work: ns	Technostress at work: +/Social support at work: ns	Psychological distress: +/Work exhaustion: ns/Perceived loneliness: +/Self-regulation: -	Age:—/gender: +/marital status: ns/education: ns/occupational area: ns/income (high): ns	Good	Direct effect	Non-significant effect of teleworking
Yeves et al., 2022 [62]	Chile	Workers in general	209	Mental health: MHI-5		Work Schedule Flexibility (WSF) (T1): ns/Work schedule flexibility (WSF) x Work from Home (WFH) (T1): ns	Work overload (T2): -	Mental health (T1): +	Gender:—/Age: c/Number of economically dependent family members: c	Fair	Moderating effect	Non-significant effect of teleworking
Kaltiainen & Hakanen, 2023 [63]	Finland	Finnish working-age population (18–65 years old) from a range of industries	996	Burnout: Burnout Assessment Tool (BAT 23 items) (T1-T2)		Increase in teleworking → Increase in job control → burnout (T1-T2):—/Increase in teleworking → Loss of social support → Burnout (T1-T2): +/increase in teleworking → Work-non-work interference → Burnout (T1-T2): +	Increase in teleworking (T2) → increase in job control → Work engagement (T1-T2): +/Increase in teleworking → increase in job control → job boredom (T1-T2):—/Increase in teleworking → Loss of social support → Work engagement (T1-T2):—/Increase in teleworking → Loss of social support → Job boredom (T1-T2): +	Increase in teleworking → Work-non-work interference → Work engagement (T1-T2):—/Increase in teleworking → Work-non-work interference → Job boredom (T1-T2): +	Having children who live at home	Good	Indirect effect of teleworking	Favorable effect of teleworking through job control and unfavorable effect of teleworking through low social support and work-non-work interferences
Neidlinger et al., 2022 [64]	Germany	German-speaking employees with leadership responsibilities and working a full-time position	367	Health: COPSOQ		Working from home intensity (WfH) (T1): +/WfH x Core Self Evaluation:—/WfH x Autonomy: ns	Autonomy (T1): +/Core Self-Evaluations (T1): +	Work life Balance (T2): WfH → Work life Balance: ns/autonomy → work-life balance: +/Core Self-Evaluation → work life Balance: +/WfH x CSE:—→ work-life balance/WfH x Autonomy: ns → work-life balance	Age:—/Gender: -	Fair	Direct and moderating effect of teleworking	Generally favorable effect

For each independent variable (factor) examined, (+) indicates increased chances of causing health problems; (-) indicates reduced chances of causing health problems or (ns) indicates a non-significant association; x indicates interaction, → indicates relationship tested in a structural equation model; n/a: non available

The overall methodological quality of the included studies was acceptable to high. Of the 20 studies, all were rated as either fair or good. No studies were excluded due to poor methodological quality. Detailed quality appraisal scores for each article are presented in Table 2.

Table 2.

Summary of the identified areas using NIH quality assessment tool for observational cohort and cross-sectional studies

Risk of bias	1	2	3	4	5	6	7	8	9	10	11	12	13	14	Score	Quality rating
Author year (Reference #)
De Carlo et al., 2022 [46]	Yes	NR	NR	NR	No	No	Yes	No	Yes	No	Yes	Yes	NR	Yes	8	Fair
Perry et al., 2018 [55]	Yes	NR	CD	CD	Yes	No	Yes	No	Yes	No	Yes	Yes	CD	No	8	Fair
Shimura et al., 2021 [57]	Yes	No	CD	CD	Yes	No	Yes	No	Yes	Yes	Yes	Yes	No	Yes	10	Fair
Van Zoonen & Sivunen, 2022 [59]	Yes	No	No	NR	Yes	No	Yes	Yes	Yes	Yes	CD	Yes	No	Yes	10	Fair
Kröner & Müller, 2023 [51]	Yes	No	CD	CD	Yes	No	Yes	No	Yes	No	Yes	Yes	CD	Yes	9	Fair
Urbanavičiūtė et al., 2023 [58]	Yes	No	CD	CD	Yes	No	Yes	Yes	Yes	No	Yes	Yes	No	No	9,5	Fair
Lazauskaite-Zabielske et al., 2023 [52]	Yes	No	CD	CD	Yes	No	Yes	No	Yes	Yes	Yes	Yes	No	No	9,5	Fair
Vanderstukken et al., 2022 [60]	Yes	No	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes	12,5	Good
Meyer et al., 2021 [53]	Yes	No	CD	CD	Yes	No	Yes	No	Yes	Yes	Yes	Yes	No	Yes	10	Fair
Kausto et al., 2024 [50]	Yes	No	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	No	Yes	No	Yes	12	Good
Galliker et al., 2024 [49]	Yes	No	CD	Yes	Yes	No	Yes	No	Yes	Yes	Yes	Yes	No	Yes	11	Good
DeWind et al., 2021 [47]	Yes	No	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes	12,5	Good
Dragano et al., 2022 [48]	Yes	No	No	Yes	Yes	No	Yes	No	Yes	No	No	Yes	Yes	Yes	11	Good
Yeves et al., 2022 [62]	Yes	No	CD	Yes	Yes	No	Yes	No	Yes	Yes	Yes	Yes	No	No	10,5	Fair
Windeler et al., 2017 [61]	Yes	No	No	CD	No	No	Yes	No	No	No	Yes	Yes	CD	Yes	8,5	Fair
Savolainen et al., 2021 [56]	Yes	No	No	CD	Yes	No	Yes	No	No	Yes	Yes	Yes	No	Yes	10	Fair
Celuch et al., 2024 [45]	Yes	No	No	Yes	Yes	No	Yes	No	Yes	Yes	Yes	Yes	No	Yes	11,5	Good
Oksanen et al., 2022 [54]	Yes	No	No	Yes	Yes	No	Yes	No	Yes	No	Yes	Yes	No	Yes	11	Good
Kaltiainen & Hakanen, 2023 [63]	Yes	No	No	Yes	Yes	No	Yes	No	Yes	No	Yes	CD	No	Yes	10	Good
Neidlinger et al., 2022 [64]	Yes	No	CD	Yes	Yes	No	CD	No	Yes	No	Yes	CD	CD	Yes	8	Fair

Yes = 1//No = 0.5//Not reported (NR) & Cannot determine (CD) = 0; Range 11–14 (Good), 7,5–10,5 (Fair), 0–7 (Poor)

The columns listed below show the risk of bias areas that we identified

Column 1. Was the research question or objective in this paper clearly stated?

Column 2. Were eligibility/selection criteria for the study population prespecified and clearly described?

Column 3. Were the participants in the study representative of those who would be eligible for the test/service/intervention in the general or clinical population of interest?

Column 4. Were all eligible participants that met the prespecified entry criteria enrolled?

Column 5. Was the sample size sufficiently large to provide confidence in the findings?

Column 6. For the analyses in this paper, were the exposure(s) of interest measured prior to the outcome(s) being measured?

Column 7. Was the time frame sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed?

Column 8. For exposures that can vary in amount or level, did the study examine different levels of the exposure as related to the outcome (e.g., categories of exposure, or exposure measured as continuous variable)?

Column 9. Were the exposure measures (independent variables) clearly defined, valid, reliable, and implemented consistently across all study participants?

Column 10. Was the exposure(s) assessed more than once over time?

Column 11. Were the outcome measures prespecified, clearly defined, valid, reliable, and assessed consistently across all study participants?

Column 12. Were the people assessing the outcomes blinded to the participants' exposures?

Column 13. Was the loss to follow-up after baseline 20% or less? Were those lost to follow-up accounted for in the analysis?

Column 14. Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure(s) and outcome(s)?

Teleworking contribution on health when examined alongside psychosocial work environment

Direct effects

Out of 20 studies, 14 explored the direct and additive effect of teleworking on health, above and beyond psychosocial work environment factors. Examining such direct effects is particularly relevant, as psychosocial work environment factors are known to account for a substantial share of variance in employee health outcomes. Demonstrating a significant direct association would therefore suggest that teleworking contributes to health beyond these established associations. Some studies found that teleworking had no significant direct effect on health (i.e., exhaustion, psychological distress, well-being) [45, 51, 55]. A study by Galliker et al. [49] also showed no significant effect of teleworking intensity on all health outcomes (i.e., depressive symptoms, somatic complaints, subjective health, emotional exhaustion). In addition, de Wind et al. [47] found that a mismatch between employees’ need for working from home and their actual access to it—that is, the discrepancy between need and access scores—was not associated with changes in fatigue one year later. Taken together, these null findings were reported in 5 out of 14 studies examining direct effects.

Despite these findings, most studies included indicated that teleworking tends to have adverse effects on health outcomes. Specifically, 7 out of the 14 studies reported predominantly negative associations between teleworking and at least one health outcome. For example, working from home predicted higher levels of depressive symptoms and anxiety [48]. Another study [54] also found that remote work predicted higher levels of COVID-19–related anxiety at the between-person level, but not at the within-person level. Similarly, Meyer et al. [53] reported that employees working entirely from home showed higher levels of exhaustion compared to those who worked part-time from home or primarily at the office. When examined by gender, this pattern was significant only among women. Specifically, women working remotely experienced heightened exhaustion during periods of strict lockdown, characterized by measures such as childcare facility closures, social distancing, and mandatory mask-wearing. Importantly, these gender differences were derived from stratified analyses rather than formal interaction (moderation) tests and therefore reflect descriptive subgroup differences rather than statistically tested moderation effects. Additional factors, including caring for young children, limited job autonomy, and insufficient partner support, further intensified this gender-specific rise in exhaustion [53].

Other research also found that the negative association between remote work and mental health outcomes may be context dependent. For instance, Savolainen et al. [56] observed that remote work predicted higher COVID-19 anxiety when controlling only for gender and age; however, this association became non-significant once psychosocial work environment factors were included in the full model. This suggests that the relationship between remote work and COVID-19 anxiety may be explained by the psychosocial work environment rather than by teleworking itself.

Another study found that working from home predicted a “suboptimal work ability” profile that decreased over time, relative to the “consistent optimal work ability” profile, assumed to be the healthiest one [50]. In this suboptimal profile, job strain was the most pronounced among the three identified profiles. Furthermore, all psychosocial work environment factors were present, which helps identify the characteristics of workers most likely to be classified within the suboptimal profile.

Taken together, these findings suggest that working from home may be associated with a less favorable trajectory of work ability over time, especially in the presence of high psychosocial risks. While working from home can offer benefits, it may also foster conditions that erode employees’ capacity to maintain optimal functioning over time.

Conversely, some research suggests that teleworking can benefit employee health. For example, working from home has been shown to be positively associated with overall health [64]. Similarly, Shimura et al. [57] found that teleworking was linked to lower psychological and physical stress responses, regardless of whether employees worked from home one or several days per week. This relationship remained significant even after accounting for increases in job-related stressors, declines in social support, greater sleep disturbances, and longer sleep duration on workdays. Overall, positive associations between teleworking and health outcomes were reported in 2 of the 14 studies examining direct effects.

Additional evidence highlights that the effects of teleworking may depend on individual or contextual factors, such as voluntariness, teleworking intensity, motivations for telecommuting, or boundary management strategies. For instance, Vanderstukken et al. [60] identified distinct profiles of telecommuters that predicted emotional exhaustion. Employees categorized in the efficiency profile—those who telecommuted to meet deadlines and manage work pressure—and in the work–life balance profile—those who telecommuted to maintain equilibrium between professional and personal life—reported lower levels of emotional exhaustion. Conversely, employees in the job requirement profile, who worked remotely primarily due to organizational necessity, did not differ significantly in their exhaustion levels. The authors also noted that members of the same telecommuting profile tended to share similar job contexts, particularly regarding work pace and autonomy.

Similarly, Urbanavičiūtė et al. [58] identified four latent boundary management profiles, each exhibiting distinct associations with exhaustion. These profiles were as follows: home-protective (rigid boundaries around the home domain and flexible boundaries around the work domain), work-protective (flexible home boundaries but rigid work boundaries), segregative (strict separation with rigid boundaries in both domains), and integrative (flexible boundaries in both home and work domains). Results showed that employees who teleworked more days per week were less likely to belong to the segregative profile. This finding suggests that more frequent telework may enhance flexibility in managing home and work boundaries, thereby reducing the tendency to rigidly separate the two spheres. Moreover, belonging to the work-protective profile at Time 1 predicted lower exhaustion at Time 2, whereas membership in the segregative profile at Time 1 predicted higher exhaustion at Time 2. These results indicate that maintaining some flexibility in home boundaries—while preserving structure around work boundaries—may protect against exhaustion over time, whereas rigid separation between home and work domains may increase vulnerability to exhaustion. By inference, these findings suggest that teleworking more days could indirectly benefit employees’ mental health by fostering more adaptive boundary management strategies. No significant effects were observed for the home-protective or integrative profiles.

Moderation effects

Beyond its direct effects, prior research has also explored how teleworking may moderate relationships between work and other characteristics and health, revealing a complex and sometimes contradictory pattern of findings. Out of 20 studies, 7 have explored the moderating effect of teleworking on health, above and beyond psychosocial work environment factors. In these studies, moderation was operationalized in two main ways: (i) teleworking as a contextual condition that alters the strength of associations between work-related factors and health outcomes, and (ii) individual or contextual factors that shape the strength of the relationship between teleworking and health outcomes.

For instance, De Carlo et al. [46] found that smart working (e.g., teleworking) amplified the positive, though initially nonsignificant, association between workload at Time 1 and exhaustion at Time 2, such that the association reached statistical significance among remote workers but not among those working on-site. Conversely, the interaction between social support and smart working was significant: teleworking attenuated the protective effect of social support, such that the negative association between social support at Time 1 and exhaustion at Time 2 remained significant only for in-person workers. This means that employees who teleworked benefited less from the buffering role of social support in preventing exhaustion. In other words, although social support generally helps reduce exhaustion, its beneficial impact appears to be diminished when employees work away from the physical workplace, possibly because teleworking limits spontaneous interactions and access to informal support networks. These results illustrate teleworking functioning as a contextual moderating condition influencing work–health relationships.

Similarly, Neidlinger et al. [64] found that the interaction between working from home and core self-evaluations had a positive impact on overall health. Specifically, the beneficial effect of working from home was more pronounced for individuals with lower core self-evaluations, suggesting that remote work may be particularly advantageous for employees with fewer internal resources. In this case, teleworking is conceptualized as the focal exposure whose effect on health varies according to individual differences.

In addition, Lazauskaitė-Zabielskė et al. [52] reported that an overwork climate predicted higher exhaustion through reduced psychological detachment, and this indirect effect was stronger among full-time teleworkers. Specifically, the mediation pathway linking overwork climate to exhaustion via psychological detachment was non-significant for part-time teleworkers. These findings suggest that teleworking may play a detrimental moderating role in this relationship, amplifying the negative impact of overwork norms. In other words, teleworkers, particularly those working full-time remotely, may be more sensitive to overwork expectations and less able to psychologically detach from work.

Other studies have identified conditions under which not teleworking appears more beneficial. For example, Yeves et al. [62] found that work schedule flexibility had a positive effect on mental health over time by reducing work overload, but only among employees who were not working from home. This pattern suggests a beneficial moderating effect of on-site work, as the positive association between work schedule flexibility and mental health was non-significant for employees working from home. The authors explained that these results may stem from the fact that teleworking during the pandemic was mandatory and implemented hastily, without sufficient preparation or organizational support.

In contrast, research on part-time teleworking (PTT) has revealed more beneficial moderation effects, particularly regarding how social job characteristics (i.e., interpersonal interaction, interdependence, and external interaction) influence exhaustion following PTT as opposed to before PTT. One study found that PTT weakened the positive link between interpersonal interaction (i.e., the extent to which a job provides opportunities to interact and engage with others) [65] and work exhaustion, yet strengthened the positive association between external interaction (i.e., the extent to which a job requires interaction with stakeholders outside one’s business unit) [65] and exhaustion [61]. However, no significant difference was observed in the relationship between interdependence (i.e., the degree to which a job depends on others to complete work) [65] and exhaustion.

These findings suggest that PTT may facilitate recovery from interpersonal interactions but not from external interactions, indeed, when external interactions increased following teleworking, levels of exhaustion tended to rise. A second study in the same paper further showed that the quality of social interactions reduced work exhaustion for both teleworkers and on-site employees, whereas a high quantity of interactions increased exhaustion. Notably, PTT mitigated this latter effect, buffering employees from exhaustion associated with frequent interactions. However, PTT also intensified the negative impact of external interactions on exhaustion, underscoring its dual role. Taken together, these findings suggest that part-time teleworking can serve as a valuable strategy for regulating the impact of social job characteristics on work exhaustion. Although interpersonal interactions are often beneficial and even rewarding, they consume psychological and cognitive energy. When individuals are exposed to a constant flow of interactions, they have fewer opportunities to halt resource depletion and to recover or build new resources.

Along the same lines, Krӧner et al. [51] found that telework exerted a beneficial moderating effect on the relationship between work intensification and well-being. Using the high well-being group as a reference, a significant interaction emerged for employees in the medium well-being group: the negative impact of work intensification on well-being was weaker when teleworking. This effect was not significant among those in the low well-being group. The same study further reported that the interaction between occupational self-efficacy (an individual resource) and telework was significantly related to the likelihood of belonging to the low well-being group—indicating that telework may offset low self-efficacy in maintaining well-being. In contrast, interactions between telework and job resources (e.g., social support, decision-making autonomy, or flexitime) were not significant. Overall, these results suggest that access to telework can enhance well-being outcomes by buffering high work intensification and compensating for low individual resources such as occupational self-efficacy, particularly in times of heightened job demands like the pandemic.

Finally, some research has found no significant moderation effects. For example, Perry et al. [55] reported that the interactions of teleworking with autonomy, emotional stability, or their combination did not significantly predict exhaustion. Similarly, Neidlinger et al. [64] observed that the interaction between working from home and autonomy did not significantly influence overall health, and this relationship was not stronger for individuals with lower levels of autonomy. Overall, moderation effects in the literature reflect heterogeneous analytical approaches in which teleworking alternates between being a focal exposure and a contextual condition within interaction models.

Mediation effects

In addition to the moderating effects discussed above, other research (n = 2/20) has examined the mediating mechanisms through which teleworking may influence employee health. For example, Van Zoonen et al. [59] found that teleworking at Time 1 increased psychological distress at Time 3 through heightened feelings of isolation at Time 2. This suggests that teleworking can foster social isolation, which in turn contributes to greater psychological distress over time.

Similarly, Kaltiainen et al. [63] reported that increases in teleworking indirectly predicted lower changes in burnout (Δ burnout) through enhanced job control. However, the same study also showed that increases in teleworking predicted higher Δ burnout through reduced social support and greater work–nonwork interference, particularly in contexts where teleworking was mandatory rather than flexible or self-selected. According to the authors, mandatory teleworking may be more disruptive to social relationships and more likely to blur the boundaries between work and personal life.

Taken together, the evidence on teleworking and employee health presents a complex and nuanced picture. Regarding direct effects, some studies found no significant impact of teleworking on health outcomes, including exhaustion, fatigue, psychological distress, and overall well-being [45, 47, 49, 51, 55]. However, most studies reported adverse effects, such as higher levels of depressive symptoms, anxiety, and exhaustion, especially among women and in contexts of strict lockdowns or limited job resources [48, 53, 54, 56]. Conversely, teleworking can also confer benefits, including lower psychological and physical stress responses and improved overall health, particularly when employees engage in flexible boundary management or telecommute for self-selected reasons [57, 58, 60, 64].

Beyond direct effects, teleworking may also moderate the relationship between work characteristics and health. Studies indicate that teleworking can attenuate or amplify the effects of workload, social support, work intensification, and individual resources on exhaustion and well-being, depending on factors such as teleworking type (full-time vs. part-time), individual resources (e.g., core self-evaluations, occupational self-efficacy), and psychosocial work environment factors [46, 51, 52, 61, 62, 64]. At the same time, some studies found no significant moderation effects, suggesting that teleworking does not uniformly interact with all psychosocial work environment or individual factors [55, 64].

Overall, research has identified several mediating mechanisms through which teleworking influences health outcomes. Teleworking may increase psychological distress through heightened social isolation [59] or indirectly contribute to burnout via changes in job control, social support, or work–nonwork interference, particularly under mandatory teleworking conditions [63]. Together, these findings highlight the importance of context, choice, and boundary management in shaping the psychological effects of teleworking.

Finally, consideration of both the voluntariness and the intensity of teleworking is crucial, given their demonstrated influence on health outcomes and their interaction with broader contextual factors (e.g., pandemic vs. non‑pandemic environments).Evidence on voluntariness suggests that employees who telework for self-selected reasons, such as efficiency or work-life balance, tend to experience lower exhaustion and more adaptive outcomes than those teleworking primarily due to organizational requirements [60]. Regarding intensity, Urbanavičiūtė et al. [58] found that employees who teleworked more days per week were less likely to belong to the segregative profile, characterized by rigid separation of home and work domains, while membership in the work-protective profile predicted lower exhaustion over time. By inference, more frequent teleworking may indirectly benefit mental health by fostering adaptive boundary management strategies. Conversely, Meyer et al. [53] observed that full-time teleworkers reported higher exhaustion than part-time teleworkers or those working primarily on-site, highlighting potential negative effects of high telework intensity. Evidence from other studies is mixed: Neidlinger et al. [64] found that teleworking intensity was favorable to overall health, including physical health, whereas other studies [49, 57] reported inconclusive effects on mental health. Overall, these findings suggest that both the choice to telework and the number of teleworking days play crucial roles in shaping psychological outcomes, yet systematic investigation of these factors remains limited. Further research is needed to clarify how voluntariness and intensity interact with psychosocial work environment factors and individual resources to influence employee health.

In sum, the literature indicates that teleworking is neither uniformly beneficial nor harmful. Its impact on employee health appears to be contingent on contextual factors, the psychosocial work environment, and individual characteristics. These findings provide a foundation for the subsequent discussion, which considers potential mechanisms, practical implications and recommendations.

Discussion

Overall contribution of the review

This systematic review aimed to synthesize longitudinal evidence on the relationship between teleworking and worker’ health, with a specific focus on the role of the psychosocial work environment and on the different analytical roles assigned to teleworking in these relationships (i.e., direct, mediation, and moderation effects). A total of 20 longitudinal studies were included, eight of which were rated as having good methodological quality. Overall, the findings highlight a still limited and fragmented understanding of how teleworking affects workers’ health when the psychosocial work environment is considered. This suggests that current evidence remains insufficiently integrated to draw definitive conclusions about causal mechanisms. This review confirms the importance of considering the psychosocial work environment as a central contextual framework in understanding teleworking–health relationships, consistent with major occupational health theories and frameworks [29–33].

Overall pattern of findings

Although the included longitudinal studies provide valuable insights, no clear or consistent pattern emerges regarding the effects of teleworking on health. This heterogeneity suggests that teleworking is not inherently beneficial or detrimental, but rather that its effects depend on a range of contextual and individual factors. In other words, teleworking effects appear to be contingent rather than universal. Moreover, the strong focus on mental health outcomes, with limited attention to physical health, further constrains a comprehensive synthesis of teleworking’s overall impact on workers’ health.

Role of psychosocial work environment

The findings underscore the central role of the psychosocial work environment in shaping the relationship between teleworking and health, although its effects remain heterogeneous across studies. Notably, more research has focused on the moderating role of teleworking, whereas indirect effects via mediators have been less frequently investigated. The 20 longitudinal studies considered well-documented psychosocial work environment factors aligned with predominant theoretical model [29–33], for example, workload, social support, job insecurity, and autonomy, as well as organisational justice as an emerging risk factor [66], along with other factors likely to arise in the evolving context of teleworking, such as flexible working hours, obligation to respond, and workplace cyberbullying. However, most of these factors were examined in a limited number of studies, which restricts the ability to draw robust conclusions about their role. All longitudinal studies that examined moderating or mediating effects did so using variables related to the psychosocial work environment. Overall, this indicates that the field remains fragmented and lacks integrative models capturing the full complexity of teleworking conditions.

Teleworking and individual resources

Additionally, three studies explored interactions between teleworking and individual resources, specifically core self-evaluation, emotional stability, and occupational self-efficacy, suggesting that individual resources may play a protective role in shaping teleworking-related health outcomes. This pattern points toward a potential buffering role of individual resources in teleworking contexts, although evidence remains limited. Across studies, additional stressors and resources were considered, with stressors generally exacerbating health problems and resources mitigating them.

Comparison with previous literature

These findings are consistent with prior systematic reviews on teleworking and health (e.g., [27, 28]), which also report heterogeneous effects. However, the present review extends this literature by explicitly integrating the psychosocial work environment as a structuring analytical framework and by focusing exclusively on longitudinal studies, thereby providing stronger insights into temporal and mechanistic processes (direct, moderating, and mediating effects).

Evidence gaps and theoretical implications

Overall, the evidence base remains fragmented and inconclusive when psychosocial work environment factors are considered. The lack of integrative models accounting for both individual and environmental variables constrained our understanding of the underlying mechanisms. These limitations emerged from the synthesis of the empirical evidence and are not intended as standalone methodological recommendations, but rather as implications derived from observed gaps in longitudinal research. There is a pressing need for theory-driven, methodologically robust longitudinal studies to elucidate causal pathways and identify key risk and protective factors. Advancing this research agenda is essential to inform evidence-based organizational and public policies that support worker’ health in remote and hybrid work settings.

Temporal gap (COVID-19 phases)

Furthermore, our review reveals that none of the included studies examined the evolution of teleworking’s health effects across distinct phases of the COVID-19 pandemic (i.e., before, during, and after). This gap in the literature limits our understanding of how shifting public health measures, organizational adaptations, and employee experiences may have shaped outcomes over time. Future research should address this by adopting longitudinal designs that capture these transitions, offering a more comprehensive view of teleworking’s impact across pandemic stages.

Quality-based interpretation

Considering study quality provides additional insight into the interpretation of findings. Overall, studies rated as higher quality tend to report either non-significant or slightly unfavorable direct effects of teleworking on health. This pattern suggests that more rigorous designs may temper overly positive interpretations of teleworking effects, while also highlighting the importance of accounting for psychosocial work environment factors. However, even in these studies, interactions between teleworking and psychosocial factors were rarely tested, limiting a full understanding of the underlying mechanisms. This represents an important methodological gap in the literature, as such interactions are likely central to understanding teleworking effects.

Synthesis by study quality

In contrast, studies rated as fair quality show highly heterogeneous and context-dependent findings, with both beneficial and detrimental effects reported. This variability further reinforces the idea that teleworking outcomes are contingent upon contextual and individual conditions rather than reflecting a uniform effect.

Final synthesis

In sum, although no consistent trend emerges across studies, there is a slight tendency among good-quality studies toward non-significant or slightly unfavorable direct effects, while fair-quality studies show even more heterogeneous results. Importantly, the limited examination of interactions with psychosocial work environment factors across studies restricts conclusions about the conditions under which teleworking may be beneficial or detrimental. Taken together, these findings highlight the absence of stable effects of teleworking on health outcomes.

Limitations

Despite the informative nature of this review, several methodological limitations should be acknowledged. First, the inclusion criteria were restricted to peer-reviewed studies published in English, excluding grey literature. Nevertheless, the confirmatory search strategy (i.e., backward and forward citation searching, or snowballing) employed was designed to be both exhaustive and systematic, reducing the likelihood of major omissions. Second, the diversity of study designs introduced challenges in terms of comparability. This issue was partially mitigated using the National Heart, Lung, and Blood Institute (NHLBI) appraisal tool [44], which provided a structured framework for evaluating methodological quality. Third, the considerable variability in follow-up durations across studies, ranging from one week to 84 months, complicates interpretation. Fourth, reliance on self-reported measures introduces potential bias. Fifth, heterogeneity in teleworking exposure measurement limits comparability. Sixth, teleworking is not a uniform experience across occupations and contexts. Seventh, although the conceptualization of the psychosocial work environment was grounded in established WHO frameworks, alternative conceptualizations exist in the literature. Consequently, our definition may have influenced which psychosocial factors and studies were included in the review, as well as the interpretation of the findings. Eighth, it should be noted that the included studies were conducted before, during, or shortly after the COVID-19 pandemic, a period characterized by exceptional and often mandatory teleworking conditions. As such, the observed associations may partly reflect pandemic-specific organizational constraints, abrupt transitions to remote work, and limited preparation or support structures. This context may differ from more recent teleworking arrangements, where hybrid work and more structured teleworking policies are increasingly implemented, potentially altering the relationship between teleworking, psychosocial work environment factors, and health outcomes. Finally, while longitudinal designs strengthen temporal ordering and reduce some limitations of cross-sectional research, they do not in themselves establish causal relationships. As recent methodological work has emphasized, longitudinal associations may still reflect unmeasured confounding, reverse causality, or reciprocal processes [67]. In occupational health research, reciprocal effects between job stressors and burnout have also been well documented, highlighting bidirectional dynamics rather than unidirectional causal pathways [68]. Therefore, the findings of this review should be interpreted in terms of temporal associations rather than causal effects.

Implications for practice and recommendations

These findings highlight several implications for occupational health practice and policy related to teleworking arrangements. Overall, these results should be interpreted within broader OHS approaches focused on the prevention and management of psychosocial risks in evolving work arrangements, including remote and hybrid work [69].

First, the absence of a consistent pattern across studies suggests that teleworking should not be conceptualized as a uniform organizational intervention. Rather, it appears to be a context-dependent work arrangement whose effects vary according to individual and organizational conditions. This supports the need for flexible and tailored approaches rather than standardized teleworking policies.

Second, in line with established occupational health frameworks (e.g., the Job Demand–Control–Support model, the Effort–Reward Imbalance model, and the Job Demands–Resources model), psychosocial work environment factors remain key dimensions systematically examined in relation to workers’ health in teleworking contexts. However, the reviewed evidence does not allow for clear conclusions regarding their role in explaining teleworking–health relationships, highlighting the need to further clarify how these factors operate in remote and hybrid settings.

Third, teleworking may be considered within OHS approaches as a work arrangement that modifies exposure to psychosocial demands and access to resources. The evidence suggests that it may reduce some stressors (e.g., commuting demands, interpersonal strain) while also potentially limiting access to social and organizational resources depending on the context. This duality aligns with the heterogeneous findings observed across studies and underscores the importance of work design considerations in teleworking implementation.

Overall, these findings support integrating teleworking into broader psychosocial risk management strategies rather than treating it as a standalone organizational practice. Policymakers and organizations should therefore adopt adaptive and context-sensitive approaches that account for both work organization factors and individual resources when designing teleworking arrangements.

Conclusions

This systematic review reveals heterogeneous evidence base regarding the health effects of teleworking when psychosocial work factors are considered. Overall, no consistent pattern emerges, suggesting that teleworking is neither inherently beneficial nor detrimental to worker’ health. Rather, its effects appear to depend on a combination of methodological, contextual, and individual factors.

Despite this heterogeneity, several psychosocial work environment factors particularly autonomy, workload, job insecurity, and social support, emerge as central determinants of health outcomes in teleworking contexts. Moreover, individual resources such as emotional stability and occupational self-efficacy appeared to moderate some effects, suggesting that teleworking interacts with both individual and environmental factors in complex ways.

These findings have important implications for the design and implementation of teleworking policies. First, given the variability in how teleworking affects individuals, organizations should avoid uniform approaches. Flexible and individualized arrangements, such as hybrid models or adjustable schedules, may better accommodate diverse needs. Second, supporting key psychosocial work environment factors remotely, training managers in virtual leadership, and leveraging teleworking as a recovery mechanism from intensive social interaction are practical strategies that can enhance health. Finally, future research should prioritize theory-driven longitudinal designs that better capture temporal dynamics (e.g., across pandemic phases), while integrating more standardized measures of teleworking exposure and objective health indicators to strengthen the evidence base.

Abbreviations

COVID-19

Coronavirus disease 2019

CINAHL

Cumulative Index to Nursing and Allied Health Literature

NHLBI

National Heart, Lung, and Blood Institute

OECD

Organisation for Economic Co-operation and Development

OHS

Occupational Health and Safety

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

PROSPERO

International Prospective Register of Systematic Reviews

PSR

Psychosocial risk

WHO

World Health Organization

Authors’ contributions

NB and APL conceptualized the study and developed the research question and overall review design. MEB, NB, and OR developed the final search strategy, including controlled vocabulary, free‑text terms, and database‑specific syntax. MEB and OR executed all database searches and documented the complete strategies in Supplementary File 1, and OR conducted backward and forward citation searches. MEB and OR independently screened titles and abstracts, while NB, OR, MEB, and APL jointly resolved uncertainties at full‑text screening and confirmed final study eligibility. MEB led data extraction and data management in Covidence, with OR contributing to initial screening datasets and NB verifying extractions during full‑text review. OR and MEB performed the qualitative synthesis of all included studies, and APL, JDG, WM, MCA, and NC verified the results of included studies. MEB and NB independently assessed study quality using the NHLBI Quality Assessment Tool, with APL resolving discrepancies. APL drafted the manuscript and conducted the revisions. NB, MEB, MCA, NC, and JG critically reviewed and edited both the original manuscript and the revisions. All authors meet the authorship criteria and have approved the final version of the manuscript for submission.

Funding

This study was funded by the Institut de recherche Robert-Sauvé en santé et en sécurité du travail (IRSST) (Grant number: IRSST 2024–0004). The funder had no role in the study design, data collection, data analysis, interpretation of results, or manuscript preparation.

Data availability

This study is a systematic review and did not generate any primary research data. The full search strategy used to identify eligible studies is provided in Supplementary File 1. The list of included studies is available in the Supplementary Materials. The review protocol was prospectively registered in PROSPERO under the registration number CRD42024599518 (https://www.crd.york.ac.uk/PROSPERO/view/CRD42024599518). All data supporting the findings of this review are derived from previously published sources.

Declarations

Ethics approval and consent to participate

Not applicable. This study is a systematic review based exclusively on previously published studies and does not involve human participants, human data, or primary data collection. Therefore, approval from an Institutional Review Board (IRB) or ethics committee was not required. As no human participants were involved, the study was not subject to the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.