Exploring the relationship between working from home, mental and physical health and wellbeing: a systematic review

Population

The population included anyone in the working population who spends all or some of their working time at home. Papers which look at students, and those studying, rather than undertaking paid employment at home, are excluded from this review. Studies which looked at the impact of temporary remote teaching on teachers (where that was not their normal mode of teaching) as a result of COVID-19 lockdown measures were also excluded from the main review (these studies are discussed separately; see Supplementary Material: Full paper excluded studies).

Exposure

This included hybrid models of home working where some time is spent working at home and some in the office or other traditional place of work. Other aspects of flexible and remote working which do not relate directly to home working, for example studies about flexible office hours or specifically about working in remote locations away from the home, along with the impact of work accessibility (e.g. the impact of remote access to emails on home life), were considered to be outside the scope of this review.

Context

The extent to which people have been asked to work at home has escalated dramatically in response to the COVID-19 pandemic and much of the recent evidence relates to the specific circumstances of home working during the pandemic. The review and model take steps to take account of this by considering evidence from both before and during the pandemic and also considering the implications for future research and policy directions.

Outcomes

Any factor that has been shown to be associated with the health of people working at home was included. An association is defined as the link between two variables (often an exposure and an outcome) where no causal relationship between the variables can be defined. This included all measures of physical health (including self-reported outcomes) and mental health (including clinical indicators such as diagnosis and treatment and/or referral for depression and anxiety alongside self-reported measures). All measures associated with wellbeing including but not limited to wellbeing, happiness, mood and stress-related outcomes were included. Work satisfaction, along with all other employment-related outcomes such as job performance and work-life balance, is outside the scope of this review.

Studies

We included quantitative, qualitative, mixed-method and observational studies. Studies with and without a comparator group were included. Books and dissertations were excluded (but references were checked for relevance in specific cases). Case studies were considered on an individual basis in terms of their study design and risk of bias. Studies from OECD4 countries only were included in the review.

Title and abstract screening

Search results were downloaded to a reference management system (EndNote). The title and abstract of each reference were screened against the inclusion criteria by one reviewer, and checked for agreement by a second reviewer. Keyword tags were used to identify the reviewer who had screened the record and to determine whether each record should be retained for consideration at the full paper stage. Full papers of potentially included studies were download as the pdf version and linked to the EndNote record for that reference. Where reviewers disagreed on the potential inclusion of a paper (i.e. one tagged the paper to be considered at the next stage and the other did not) the full paper was obtained to clarify the relevance of the work to the inclusion criteria. This was agreed by consensus between the three reviewers.

Full paper screening

The full paper of all potentially relevant papers was read by one reviewer. Where a decision to exclude the paper (due to lack of relevant data) was made the reason for this decision was tabulated and checked by a second reviewer. Uncertainties were resolved by discussion between the three reviewers and among the wider review team as required until a definitive list of included papers was obtained.

General health

Prior to the COVID-19 pandemic, the evidence base for the impact of working from home on overall health was limited. We identified only seven studies which considered these factors (Hornung et al. 2009,5 Vitterso et al. 2003,10 Virick et al. 2010,9 Kaduk et al. 2019,6 Sardeshmukh et al. 2012,8 Perry et al. 2018,7 Wöhrmann et al. (2021)11). The factors reported in these studies are summarised in Table 5. Reported factors which influence the associations between working from home and general health (pre-COVID-19 papers), and Figure 2. Mind map of factors which influence the associations between working from home and physical or general health (pre-COVID-19 papers).

FIGURE 2.

Mind map of reported factors which influence the associations between working from home and physical or general health (pre-COVID-19 papers). Key: Positive effect, Negative effect, No effect. See Table 2 for index of studies.

Two cross-sectional survey studies (Hornung et al. 20095, Vitterso et al. 200310) both demonstrated a positive effect on QoL from working at home, although factors which mediate this effect included autonomy and work–family conflict (Hornung et al. 20095), along with a sense of belonging versus isolation (Vitterso et al. 200310). The intensity of telecommuting was also considered, with hybrid working arrangements being seen to be most positive in both studies.

Hornung et al. (2009)5 conducted their study with German public administrators (including teleworkers and non-teleworkers). Their postal survey (n = 1008) had a response rate of 67% and was analysed using structural equation modelling. The workers were 27.5% female with a mean age 43.6 (SD 8.8); 62.6% had a teleworking arrangement (1–4 days per week, mean 1.7 (SD 1.1) days). QoL was assessed using the WHOQOL-BREF, and autonomy, work–family conflict and job satisfaction were also measured. There were small but statistically significant positive effects of telecommuting intensity on QoL mediated via both autonomy (β_indirect = 0.02, z = 2.56, p < 0.01) and work–family conflict (β_indirect = 0.11, z = 5.96, p < 0.01).

Vittersø et al. (2003)10 conducted their survey in Europe (UK, Norway, Iceland, Portugal) via self-administered questionnaires with a response rate of 41%. A subset of qualitative interviews on a more focused set of issues were also undertaken but there were no details on the collection or analysis of qualitative data. Workers (n = 217) were selected from companies where ‘teleworking seemed possible / likely’ (including those who did and did not telework). The sample means age was 38.25 years. Subjective QOL was assessed using the CHP’s QoL Profile, measuring four dimensions (three of human growth and one of overall life satisfaction) on a 5-point scale. Employees’ sense of belonging increased with a greater number of days working from home (β = 0.30, p < 0.001). There were no significant impacts of WFH on control, flexibility or concentration in this model. Those who had worked for a whole week at home described it as an ‘isolating experience’.

A positive effect on life satisfaction was observed with working at home by one further cross-sectional survey study, again mediated in the same way by the hybrid working approach (Virick et al. 2010)9. Virick et al. (2010)9 conducted an online survey with 85 telecommuters working for a large telecoms organisation in the USA. They were mostly male (75% male) with a mean age of 41.5 (SD 9.15) years. Life satisfaction was assessed using a four-item validated scale. The extent of telecommuting was assessed using an item asking respondents to estimate the average number of days worked from home (1–6 scale). They also assessed job satisfaction, worker types, perceived performance outcome orientation and control variables (tenure and gender). A ‘curvilinear relationship was identified’, whereby ‘for employees with high drive and low enjoyment, life satisfaction is low when there is a moderate amount of telecommuting. However, with other employees, the relation is the opposite. Life satisfaction is high when there is a moderate amount of telecommuting’. Therefore drive and enjoyment moderate the relationship between telecommuting and life satisfaction.

Exhaustion, strain and the risk of burnout were reported as negative effects of working at home by three cross-sectional surveys, one of which also included a longitudinal element for some participants (Kaduk et al. 2019,6 Sardeshmukh et al. 2012,8 Perry et al. 20187) The associations were made worse by high job demands (Sardeshmukh et al. 20128). Importantly, Kaduk et al. (2019)6 were the only authors to consider the difference between voluntary and involuntary working from home in the pre-COVID era. In addition, Wöhrmann et al. (2021)11 found no significant correlations between telework and psychosomatic health complaints. Their study was based on data collected in 2015.

Kaduk et al. (2019)6 conducted an in-person survey of IT workers (Fortune 500 companies) in the US (n = 758) with a 69.6% response rate. Six wellbeing outcomes of work–family conflict, job satisfaction, turnover intentions, emotional exhaustion (burnout), perceived stress and psychological distress were considered. Over 95% of employees reported doing some work at home; 31% do at least 20% of working time at home (‘substantial remote work’) voluntarily and 14% involuntarily. They reported that involuntary variable schedules are associated with greater work-to-family conflict, stress, burnout and turnover intentions, and lower job satisfaction in models that adjust for personal characteristics, job, work hours, family demands and other factors. Voluntary remote work was found to be protective, and associated with less stress (although this relationship was attenuated in the full models).

Sardeshmukh et al. (2012)8 surveyed telecommuters working for a large supply-chain-management company in the Midwestern USA (n = 417) with a 37.9% response rate. The sample were 29% female, with mean age 26–35 years. They had spent at least 1 year teleworking on average, spending 8–40 hours a week teleworking and most telecommuting less than 4 days per week. Exhaustion was assessed using the Maslach and Jackson validated scale. They also assessed time pressure, role ambiguity, role conflict, autonomy, feedback, social support and job engagement. They found that telework is negatively related to both exhaustion and job engagement. Job demands and resources (time pressure, role ambiguity, role conflict) partially mediated the relationship between the extent of telework and exhaustion.

Wöhrmann et al. (2021)11 reported on a telephone survey (described as large-scale, population-level) undertaken in 2015 in Germany with highly qualified white-collar workers aged below 65 years who reported the use of modern information and communication technology for their work. The sample were 49% female, with mean age 46.2 (SD 10.2) years. Of these, 18% teleworked, with an average of 1.92 days per week. Psychosomatic health complaints were assessed by aggregating into an index and included headache; fatigue, weariness or lassitude, stomach and digestion complaints, tension and irritability, sleep disorders, dejection, physical exhaustion and emotional exhaustion. Participants rated whether they occurred frequently in the last 12 months on work days. No significant correlations were found between telework and psychosomatic health complaints. They note that generalisability may have been limited by the disproportionately small numbers of teleworkers in Germany at the time of the survey (in comparison with international numbers).

Perry et al. (2018)7 conducted an online cross-sectional survey (with a longitudinal second-wave survey at 3 months for some participants) of full-time professional employees from three southern USA organisations. The overall response rate was not reported. Their sample (n = 258) were 55% female, with mean age 55 years. Emotional stability was assessed by to 10-item subscale of the IPIP Big Five personality scale and strain was assessed by the Oldenburg Burnout Inventory (8 items) to measure exhaustion and disengagement. There was a significant remote work × autonomy interaction for exhaustion, such that there was ‘a positive remote work-exhaustion slope among employees reporting low autonomy (0.82; t = 2.12, p < 0.05) but no significant relationship among those reporting high autonomy (slope = −1.20; t = −1.68, p = 0.10)’ Those with high autonomy and high emotional stability exhibited the lowest overall level of strain compared with other combinations of these predictor variables.

Physical health

The effect of working at home on measures of, or with direct links to, physical health were reported in five studies prior to the COVID-19 pandemic, again, reflecting a limited evidence base for these measures. The factors reported in these studies are summarised in Table 6. Reported factors which influence the associations between working from home and physical health (pre-COVID-19 papers), and Figure 2. Mind map of reported factors which influence the associations between working from home and physical or general health (pre-COVID-19 papers).

The impact on overall physical health (using perceived health / self-reported measures) was neutral for the two studies which considered it (Kroll and Nuesch 2019,103 Mann et al. 200325). These consisted of longitudinal survey data from 1999 and 2009 (Kroll and Nuesch 2019103) and a cross-sectional mixed-methods survey which included semi-structured interviews and a survey (Mann et al. 200325).

Kroll and Nuesch (2019)103 selected a representative sample (n = 6132) of those living in private households in Germany, limited to those aged 20–60 and employees (not self-employed). The analyses of working from home are based on data from the 1999 and 2009 waves, but were analysed as cross-sectional. Perceived health was assessed by the question ‘How would you describe your current health?’, with response options being very good, good, satisfactory, poor, bad. Job satisfaction and leisure satisfaction were each assessed by the question ‘How satisfied are you today with your job/your leisure time?’ Working from home did not have a significant effect on health when controlling for individual heterogeneity (b = 0.02, SE = 0.05, ns). There was also no statistically significant effect of working from home on leisure satisfaction, however (b = −0.01, SE = 0.10, ns).

Mann et al. (2003)25 conducted a study in the USA consisting of qualitative interviews (comparing the emotional impact of work patterns on teleworking and office-based journalists) and a quantitative questionnaire (comparing the occupational stress and health symptoms of office-workers and teleworkers). The study population consisted of eight males and four female (study 1) and 17 males and 15 females (study 2). Teleworking participants followed this working pattern (minimum of 3 days per week) from 3 months to 10 years (average 3 years 2 months) out of a total average working time of 17 years 11 months. Physical health was measured by self-reported frequency of physical symptoms associated with occupational stress. Mental health was measured by the feelings and behaviours that were perceived to be affected by the pressure of the job. There was no significant difference between the physical health scores for the teleworkers and office-workers (t = 1.05, df = 60, ns).

The effect on physical inactivity was mixed, with two studies (a retrospective cohort study and a national cross-sectional travel survey) reporting a positive effect on physical inactivity including increased walking and cycling (i.e. becoming more active) (Henke et al. 201620, Chakrabarti et al. 201892) and one further qualitative (interview) study reported increased sitting as a result of working at home (Grant et al. 201318). Henke et al. (2016)20 also considered the impact on alcohol and tobacco use and reported a decrease in both as a result of working at home. This was moderated by ‘telecommute intensity’, that is, the predicted probability of being at risk appeared to increase with increasing number of telecommuting days.

Chakrabarti et al. (2018)92 conducted a cross-sectional analysis of a national dataset in the USA. Their sample consisted of employed people aged 18–64 (n = 123,810) of whom 62.1% reported telecommuting (‘working at home for an entire work day’). Physical activity (PA) was assessed using the Day Trip File (one-day travel diary) calculated into minutes of PA, assumed to be at least moderate intensity. Telecommuters were classified into ‘frequent’ (more than 4 days over the past month or once per week on average) and ‘occasional’ (1–3 days over the past month). Both frequent and occasional telecommuters engaged in 8–9 minutes more per day of PA than non-telecommuters, on average; 31% frequent, 27% occasional and 21% non-telecommuters met or exceeded the 30 minutes per day activity target.

Grant et al. (2013)18 conducted in-depth interviews (n = 11) with UK e-workers, across five organisations and three sectors. All participants worked remotely using technology independent of time and location for several years. The focus of this study was on wellbeing. However, amongst the many wellbeing outcomes measures (discussed elsewhere in this report), it was noted that sitting behaviours may increase as a result of working from home.

Henke et al. (2016)20 conducted a longitudinal cohort study (retrospective) of employee data at one USA insurance firm from 2010 to 2011 (n = 3703). The sample were 62% female, 88% aged less than 55 years, 58% were prime-time telecommuters, 20% were off-hour telecommuters and 22% were non-telecommuters. The physical health outcome considered was ‘high risk for physical inactivity’ (reported less than 3 days of cardiovascular exercise per week). In 2011 a greater proportion of non-telecommuters were at risk of obesity, poor nutrition, physical inactivity (not statistically significant) and tobacco use. ‘Prime-time telecommuters’ (described as those with at least 51% of remote hours during prime work hours of 06.00–18.00) working at least 73 hours per month (very high intensity) had a significantly lower risk of alcohol abuse compared with non-telecommuters (1.8% vs. 2.9%). Prime-time telecommuters working 9 to 32 hours per month (medium intensity) had a significantly lower risk of physical inactivity compared with non-telecommuters (35% vs. 41%). Prime-time telecommuters working 33 to 72 hours per month (high intensity) had a significantly lower tobacco risk compared with non-telecommuters (4.5% vs. 7.2%).’

Mental health and wellbeing

A more substantial volume of evidence exists which consists of 24 studies considering the effects of working at home on a broad range of wellbeing and mental health-related measures prior to the COVID-19 pandemic. The factors reported in these studies are summarised in Table 7. Reported factors which influence the associations between working from home and wellbeing and mental health (pre-COVID-19 papers) and Figure 3. Mind map of reported factors which influence the associations between working from home and wellbeing or mental health (pre-COVID-19 papers).

FIGURE 3.

Mind map of reported factors which influence the associations between working from home and wellbeing or mental health (pre-COVID-19 papers). Key: Positive effect, Negative effect, No effect. See Table 2 for index of studies.

There was no clear overall pattern to the factors which influence the associations reported between working at home and the broad range of wellbeing measures investigated by study authors prior to the COVID-19 pandemic. This lack of clarity highlights the complexity in the relationships being considered, and indicates how people are likely to experience working at home differently depending on their circumstances. A range of factors relating to overall wellbeing, depression and other mental health measures, measures of stress, control over the workspace (including ability to work undisturbed, interference with family life in the home, and fears associated with technology failure), social contact, communication with colleagues, and measures of satisfaction and fulfilment were considered. The evidence was from self-reported data collected via surveys, qualitative interview studies and retrospective cohort studies.

Wellbeing and mental health measures

Wellbeing as a result of working at home was reported by one survey to have increased overall (Hall et al. 201917) as rates of depression were reportedly reduced for some telecommuters in a further retrospective cohort study (Henke et al. 201620). Positive affect was also shown to increase with working at home and this was furthered by having better coping skills, being more open to new experience and having greater social connectedness (Anderson et al. 201413), as was a decrease in ‘psychological strain’ further enhanced by reduced social isolation (Bentley et al. 2016)14. However, Mann et al. (2003)25 conversely reported that emotional ill health increased for those working at home (survey and semi-structured interviews). Further, Mellner et al. (2017)26 found no effect on ‘psychological detachment’, and Charalampous et al. (202112) found no effect on psychosomatic health for home workers.

Hall (2019)19 in their online survey of UK employees who work two or more days a week from home (n = 897) stated that ‘homeworking can increase employee engagement, job satisfaction and wellbeing’. However, their research question was not clear and the paper contained very little detail. Henke et al. (2016)20 conducted a longitudinal cohort study (retrospective) of employee data at one USA insurance firm from 2010 to 2011 (n = 3703). The sample were 62% female, 88% aged less than 55 years, 58% were prime-time telecommuters, 20% were off-hour telecommuters and 22% were non-telecommuters. They reported that those who telecommuted for less than 8 hours per month (low-intensity telecommuters) were likely to reduce their rate of depression at a greater rate than non-telecommuters over time.

Anderson et al. (2014)13 in their survey of a ‘large US federal agency’ (n = 102) measured outcomes on the ‘Job-Related Affective Well Being Scale’. They also measured openness to experience, trait rumination and sensation-seeking, and social connectedness outside the workplace. They report that the relationship between telework and positive affect is moderated by one’s social connectedness outside of the workplace such that the relationship becomes more positive as social connectedness increases (γ = 0.75, p < 0.001); also individuals experience less negative affect while teleworking as social connectedness increases (γ = −0.73, p < 0.01). In addition, Bentley et al. (2016)14 in their online survey of ‘knowledge workers’ in New Zealand (n = 804, 47% female, mean age 30.9 (SD 11.4) found that psychological strain was significantly predicted by organisational social support (this did not differ significantly between hybrid and low telework (1–7 hours) and teleworker support (although this was only significant in the whole sample and not in the hybrid and low telework sub-samples)). Mann et al. (2003),25 described above, found there was a significant difference between the mental health scores of home and office workers f (t = 1.85, df = 60, p < 0.05), indicating higher levels of emotional ill health for the teleworkers.

Charalampous et al. 202112 conducted semi-structured interviews with 40 (23 male) remote e-workers working for a British IT company (mean age of 46.86 (SD = 8.43)) to explore the impact of the remote e-working experience on employees’ wellbeing. Regarding individuals’ psychosomatic health, they report that ‘it appeared that none of the employees reported serious and exaggerated health conditions resulting from remote e-working’. However, increased sedentary behaviours combined with the absence of breaks was a prominent risk factor for psychosomatic health. Regardless of the risk of not taking enough breaks, and not having appropriate ergonomics, interviewees suggested that a healthier lifestyle was available to them. Mellner et al. (2017)26 conducted an online survey with Swedish professionals working in one of four large organisations (n = 3846, 62% male, mean age 48 years). Working from home was negatively related with sleep duration: This is reported as follows: ‘In this context [of boundaryless working conditions], not being able to free oneself from work-related feelings and thoughts during leisure may be interpreted as the dark side of freedom as employees run the risk of working “anytime – all the time”, and as such, of “always being on” resulting in disturbed sleep.’

Stress

A significant number of studies considered how stress influences the association between working at home and health, with a similar number of studies reporting positive and negative effects. Working at home was reported to have a positive association with self-reported stress (i.e. stress was reduced as a result of working from home) in six studies (including cohort and survey and qualitative studies) (Henke et al. 2016,20 Kaduk et al. 2019,6 Mann et al. 2003,25 Song et al. 2020,30 Tietze et al. 201129). These positive relationships were further enhanced by having less office politics (Henke et al. 201620 and Mann et al. 200325), no travel to work (Mann et al. 200325), more control over work (Mann et al. 200325) and by reduced telecommute intensity (hybrid working) (Henke et al. 201620). A reduction in specific stress related to childcare and travel was also reported in one final qualitative study (Grant et al. 201318). Conversely perceived stress was negatively associated with working at home in two studies (survey and retrospective cohort) (Kaduk et al. 2019,6 Ray et al. 202127). Kaduk et al. (2019)6 reported that the negative association was made worse by working more hours, family interference in work and having a spouse in the home, being female or non-white, and having low income or low health status. Stress was linked to issues with technology by Charalampous et al. (2021)12. In addition Trent et al (1994)34 found no association between working from home and stress in their survey study.

In addition, one further observational study objectively measured levels of physiological stress response by taking blood pressure readings along with urine and saliva samples (Lundberg and Lindfors 200224). They reported no significant difference in self-rated stress between telework and office work; however, blood pressure was significantly higher during work at the office than when teleworking at home, and men had significantly elevated epinephrine levels in the evening after telework at home.

Ray et al. (2021)27 reported on a USA-based, face-to-face nationally representative survey (n = 7400, mean age 42.8 years, 52% female). The proportion working from home increased from 29% in 2002 to 33% in 2018 (remaining stable from 2010 onwards). In regression analyses, working from home was associated with a 22% increase in job stress and a 65% increase in job satisfaction (p < 0.01). This was modified by sex (women were 38% more likely to report job stress and 5% less likely to report healthy days than men), ethnicity (compared with non-Hispanic Whites, Hispanic workers were 32% less likely to report job stress, 5% more likely to report healthy days and 39% less likely to report days with activity limitations; Black workers were 40% less likely to report job stress and 4% more likely to report healthy days; and Asian workers were 36% less likely to report job stress and 4% more likely to report healthy days).

Song et al. (2020)30 conducted telephone interviews and time use survey with workers in the USA (n = 3962). Telework on weekdays or weekends/holidays was associated with more stress (p < 0.005) than working only in the office. Tietze et al. (2011)29 conducted interviews and focus groups with people working in local authorities in the UK. Pre-implementation, participants expected to gain better personal wellbeing form working from home, particularly in relation to being calmer and less stressed. Post-implementation, this expectation was realised for many participants, who felt less stressed and more relaxed. Some attributed it to escaping ‘bickering and gossiping’ in the office.

Trent et al. (1994)34 conducted a mailed survey with people working at private-sector companies in the USA (n = 38; 15 telecommuters, mean age 46.3 (SD 7.0); 9 people who worked exclusively from home, mean age 33.8 (SD 6.0); 14 people who worked exclusively from the office, mean age 41.7 (SD 8.6) years). Stress was assessed as perceived social support, assessed using the Social Support Index. There were no differences in perceived stress score between the three groups (telecommuters, those who worked from home, those who worked from the office), although the authors suggest this may be due to the sample size. Isolation scores were highest among the work-at-home group (mean 3.1, SD 1.1), higher than the office group (mean 2.4, SD 0.9), and lowest in the telecommuting group (mean 1.7, SD 1.0), and the ANOVA showed a significant difference of group (F = 5.82, p = 0.007).

Lundberg and Lindfors (2002)24 conducted a ‘repeated measures observational field study’ of white-collar workers at a Swedish government authority who worked 3 days at the office and 2 days per week at home (n = 26, 46% female, mean age 41.7 (SD 11.2) years). Psychophysiological reactivity, using an automatic ambulatory blood pressure monitor on three separate days, was undertaken. Urinary catecholamines and salivary cortisol were also measured at regular intervals during this period. Participants also reported self-rated health and wellbeing (Ryff’s Psychological Well-Being Scales). Women had significantly higher daytime systolic blood pressure (SBP) at the office than during telework or relaxation, but no differences between telework and relaxation. Men had significantly higher daytime SBP at the office than during relaxation at home, but no significant differences between office work and telework, or between telework and relaxation. Women, but not men, had significantly higher daytime DBP during work at the office than during telework. Women had significantly higher daytime epinephrine levels during telework than relaxation, and during office work than relaxation. Men had significantly higher daytime epinephrine levels during office work than relaxation, but not during telework. There were no significant effects of session for cortisol at any time period.

The remaining five studies have been described above. In terms of measures of stress: Henke et al. (2016)20 (longitudinal cohort study, n = 3703) reported that prime-time telecommuters had the highest risk of stress (but this was not statistically significant). Kaduk et al. (2019)6 (in-person survey n = 758) found that involuntary variable home working schedules were associated with greater stress. Voluntary remote work was found to be protective, and associated with less stress (although this relationships was attenuated in the full models). Mann et al. (2003)25 found a significant difference in mental health scores f (t = 1.85, df = 60, p < 0.05), indicating higher levels of emotional ill health for the teleworkers. Grant et al. (2013)18 (described above: in-depth interviews n = 11) reported that working from home can relieve stress from travel and child-care issues. Charalampous et al. (2021) 12 (semi-structured interviews, n = 40) reported that feelings of anger, frustration and stress were mainly linked to issues with technology, or not being able to get hold of colleagues when needed.

Life satisfaction

Factors which influence the associations between working at home wellbeing with measures of life satisfaction were reported in two studies. In a retrospective cohort study, Reuschke et al. (2019)28 reported no effect of working from home on life satisfaction overall, and no mitigation of that relationships by gender or type of employment. However, they did report a positive association between working at home and leisure-time satisfaction. Daniel et al. (2018) 16 reported a positive association between fulfilment and working at home, an association which was further increased by having greater autonomy at work.

Reuschke et al. (2019)28 analysed a large UK longitudinal dataset of people of working age (18–64 years) in either paid employment or self-employment (n = 33,719, 53.7% female); 11.1% (n = 3738) mainly worked from home in at least one wave. Home working was found to be positively associated with health satisfaction in men, but not in women (although this was not significant in their final model). The authors explain this in terms of both home working and self-employment having common elements, namely autonomy and control, and hypothesised that men ‘value these things’ – although this is a clearly subjective viewpoint. For women, both home working and self-employment (as an employer and solo) were significantly positively related to job satisfaction, and home working has an ‘additional’ benefit to the advantages gained from being self-employed. No data are presented in this discursive paper. Daniel et al. (2018) 16 conducted qualitative interviews with online home-based business owners in the UK (n = 23, 65% female). Participants described feeling more fulfilled by having more time and mental space for creativity and creative work, by working online and thus being ‘freed’ from daily face-to-face workplace interactions, which they saw as a distraction. Participants also enjoyed the autonomy of scheduling inherent in working from home (particularly with their own business), for example alternating working patterns to fit around preferences for each day, such as taking an extended lunch break and catching up in the evening.

Social interactions

More links to wellbeing were reported through changes to level of social interactions, with four studies reporting a negative association due to a decrease in social interaction and adult contact as a result of working at home (Charalampous et al. 2021,12 Collins et al. 2016,15 Koehne et al. 2012,23 Stitou et al. 201831). Conversely one study reported improvements in people’s social lives from working at home (Jacukowicz et al. 202022). Communication with and support from colleagues were also reported as critical success factors for working at home (Grant et al. 2013,18 Charalampous et al. 202112). This was further supported by Mann et al. (2003)25.

Collins et al. (2016)15 conducted semi-structured interviews with staff in a local authority in the UK (n = 33, demographics not reported). Teleworkers only had social relationships with office workers whom they already knew and had already met face-to-face. They did not know nor get to know new office staff. There was no social network functionality set up to facilitate interaction between teleworkers. Koehne et al. (2012)23 conducted interviews with individuals working primarily from home, or who worked in the office but had colleagues working from home in the USA, Estonia, UK, Spain and Mexico (n = 17, 29% female). They report that social isolation was an issue for 10/17 participants. Remote workers would counteract this by seeking social interaction in their home communities. Stitou et al. (2018)31 also conducted interviews with female, home-based childcare workers in Canada (n = 11). Factors affecting health and wellbeing were reported as the absence of contact with other adults during working hours, a lack of external help during working hours (i.e. working alone, without breaks), difficulty filling spots, noise, interference with personal and family life, low and precarious remuneration, and incomplete or no benefits.

Jacukowicz et al. (2020)22 conducted an online survey of traditional office workers (n = 200, 60% female) and online workers (n = 189, 82% female) in Poland. Quality of social relationships was assessed using one subscale of the WHOQOL-BREF questionnaire. Working online significantly predicted lower satisfaction with work-life balance (β = –0.17, p < 0.01) but greater quality of social life (β = 0.13, p < 0.05).

Charalampous et al. (2021)12 (semi structured interviews, n = 40) reported that numerous remote e-workers said that it was easy to feel lonely, bored and sad when the social interaction was reduced or eliminated when working at home. Grant et al. (2013)18 coducted in-depth interviews (n = 11), reporting that home working was ‘wellbeing detracting’ as social interaction may be limited to family and local friends. Mann et al. (2003)25 (qualitative interviews) reported that teleworkers emphasise the lack of social support available to talk things through which could produce other negative emotions such as feelings of insecurity and lack of confidence in their abilities.

Control

In terms of the home working environment, positive impacts on wellbeing were reported as a result of having more control over the workspace and times of work (Hislop et al. 2015,21 Mann et al. 2003,25 Thulin et al. 2019,32 Tietze et al. 201129), and having the ability to work undisturbed (Thulin et al. 201932). However, Stitou et al. (2018)31 reported an increase in interference with family life as a result of working from home. Hislop et al. (2015)21 reported a negative association with wellbeing due to a ‘sense of perpetual contact’ arising from working at home and the report of ‘self-imposed pressure to perform’. In addition, ‘fears of technical failure’ and the subsequent negative impact on wellbeing reportedly increased with working from home (Daniel et al. 2018)16.

Hislop et al. (2015)21 conducted qualitative interviews with self-employed home workers undertaking IT-based office support / administrative work in the UK (n = 14, gender/age not reported). The home workers reported being happy with their work, with the main benefit being the ‘spacio-temporal flexibility inherent in homeworking’, They could structure their own time and work location to balance their work with domestic commitments such as shopping and childcare. The negative aspects of home working included increasing people’s sense of ‘perpetual contact’, creating a sense that work was difficult to escape from.

Thulin et al. (2019)32 conducted an online survey in Sweden (n = 456, 70.6% female, mean age 43.2 (SD 10.9) years) with a response rate of 40%. In the logistic regression model for time pressure, never teleworking (β = −0.644, p < 0.05), only teleworking within regular hours (β = −0.866, p < 0.01), age (being older; β = −0.032, p < 0.01), working full time (β = −0.806, p < 0.05) and using a smartphone for private purposes often (β = −1.115, p < 0.05) or all the time (β = −1.089, p < 0.05) were associated with experiencing less time pressure, whereas having children at home (β = 0.406, p < 0.01) was associated with experiencing more time pressure.

Mann et al. (2003)25 (qualitative interviews) reported that teleworkers may experience a decrease in stress attributable to the perception of having control over their work (environment and work schedules). Tietze et al. (2011)33 (interviews and focus groups) reported that working from home made it easier for participants to manage their own workloads and consequently address equity issues and experience improved wellbeing. However, Stitou et al. (2018)31 (interviews, n = 11) reported that factors affecting health and wellbeing included interference with personal and family life. Daniel et al. (2018) 16 (qualitative interviews, n = 23) reported that fear of the IT equipment/online connection failing was ‘pervasive and driven by a fear of being completely cut off’.

General health

Even as a result of increased working from home due to COVID-19 the volume of literature linking working at home with general health outcomes has not increased substantially. A further five studies linking the outcomes of QoL, higher demands and lifestyle factors (diet and alcohol intake) show a mixed picture in terms of their impact on the associations with working at home. These associations reported in these studies are summarised in Table 8. Reported factors which influence the associations between working from home and general health (COVID-19 papers) and Figure 4. Mind map of reported factors which influence the associations between working from home and general health (COVID-19 papers).

FIGURE 4.

Mind map of reported factors which influence the associations between working from home and general health (COVID-19 papers). Key: Positive effect, Negative effect, No effect. See Table 2 for index of studies.

Quality of life was shown to be positively associated with working at home in a survey by Weitzer et al. (2021)39. Weitzer et al. (2021)39 conducted an online survey of the general population is Austria (n = 1007, 55% female, age 18–65); 29.3% were working from home part of the time and 21.0% were working from home all of the time. Those who worked from home all the time were more likely to report an increased QoL compared with those who were not working from home (OR 3.69, 95% CI 1.86 to 7.29). Working part of the time from home was also associated with an increased QoL compared with not working from home (OR 2.07, 95% CI 1.09 to 3.91). Likewise, not working from home appeared to be associated with decreased QoL compared with working part or all of the time from home. The positive association was strongest for older participants, men, and those who were university educated.

Conversely, general health was shown to be negatively associated with working from home in a prospective cohort study by Di Tecco et al. (2021)35. The association was also negatively affected by lower education level, higher work demands and poorer management of change. Di Tecco et al. (2021)35 conducted a prospective cohort study to investigate the impact of a ‘smart working pilot’ (flexible home working) in Italy (n = 187, 78.6% female, mean age 50.7 (SD 6.8) years). There was no significant change in general health (p = 1.00) or wellbeing (p = 0.247) as a result of the pilot as evaluated by t-tests. In the regression models, significant predictors of wellbeing were demands (−0.703, p = 0.027) and effective management of change (1.461, p = 0.003), and demands (−1.00, SE 0.048, p = 0.037) and higher education (0.238, SE 0.100, p = 0.018) significantly predicted general health.

In terms of lifestyle measures of health, factors which influence the associations between working from home and both an improved diet and reduced alcohol consumption were reported in a survey by Sato et al. (2021b38). These positive associations were seen more strongly in men than women. However, Kubo et al. (2021)36 reported negative associations between working from home and food behaviours including increased solitary eating and lower meal frequency (with more days of telecommuting increasing the negative associations). In considering drug use, Ripoll et al. (2021)37 found no association with working from home (survey study).

Sato et al. (2021b)38 conducted an online survey in Japan (n = 5929, 69% female, mean age 44.0 (SD 13.8)) Working from home was associated with increased intake of vegetables (1.02, 1.004–1.03), fruits (1.06, 1.03–1.09), dairy products (1.03, 1.01–1.06) and snacks (1.04, 1.02–1.06) but decreased intake of seaweeds (0.94, 0.91–0.97), meats (0.98, 0.96–0.999) and alcohol intake (0.93, 0.86–0.997). Among women, time spent on childcare was associated with reduced intakes of vegetables and fruits.

Kubo et al. (2021)36 also conducted an online survey of office workers who telecommuted in Japan (n = 13,468, 48.8% female, age 20–65). Those who telecommuted more frequently tended to have more unfavourable eating habits. Among workers who hardly telecommuted, 25.5% missed breakfast, 25.9% ate all meals alone, 1% ate less than two meals a day, and 6.6% adopted meal substitution. The corresponding proportions among workers who telecommuted in excess of four days per week were 28.7%, 37.0%, 2.5% and 8.4%, respectively. The odds ratios (95% CI) for those who telecommuted at least 4 days per week relative to those who rarely telecommuted were: skipping breakfast: 1.15 (1.03 to 1.29); solitary eating: 1.44 (1.28 to 1.63); lower meal frequency: 2.39 (1.66 to 3.44); and meal substitution: 1.26 (1.04 to 1.51).

Ripoll et al. (2021)37 conducted an online survey in Spain (n = 681, 77% female, age 18 and above). Working from home (compared with other working arrangements) was not associated with increased consumption of psychotropic drugs between weeks 1 and 4 (consumed by 6.5% and 7.1%, respectively, p = 0.306) or weeks 1 and 8 (consumed by 8.4% and 8.9%, respectively, p = 0.952), or consultations to improve mood/anxiety between weeks 1 and 4 (undertaken by 27.3% and 26.9%, respectively, p = 0.918) and weeks 1 and 8 (undertaken by 28.0% and 31.5%, respectively, p = 0.388).

Physical health

The COVID-19 pandemic has also resulted in a slight increase in the number of studies reporting factors which influence the associations between working at home and physical health measures. These factors reported in these studies are summarised in Table 9. Reported factors which influence the associations between working from home and physical health (COVID-19 papers) and Figure 5. Mind map of reported factors which influence the associations between working from home and physical health (COVID-19 papers). Notably all the factors reported had a negative effect on the relationships (or no change in the association was found). This entirely negative framing of the research is not seen in any of the other sets of data considered in this report. In total seven studies considered a range of physical health measures such as physical wellbeing, physical function, physical limitations, musculoskeletal problems, pain, PA and sedentariness and reduced work comfort.

FIGURE 5.

Mind map of reported factors which influence the associations between working from home and physical health (COVID-19 papers).

A negative association between physical wellbeing and working from home was reported in two studies (survey and qualitative) (Waizenegger et al. 202043 and Xiao et al. 202145). The negative relationship was mediated by poor mental wellbeing, low PA, low ‘healthy’ food intake and high ‘junk’ food intake (Xiao et al. 202145) along with having minimal contact with others due to the lack of sports facilities open during COVID-19 lockdowns (Waizenegger et al. 202043).

Waizenegger et al. (2020)43 conducted a ‘qualitative interpretive study’ of knowledge workers who were previously working in office spaces and worked from home during COVID-19 lockdowns in various countries (n = 33, 39.4% female, aged 20–50). They report that ‘participants’ mental and physical wellbeing was (equally) affected by a lack of PA, due to sports facilities being closed and minimal contact with others being allowed’.

Xiao et al. (2021)45 conducted an online survey in the USA of office workers who had transitioned to WFH during the COVID-19 pandemic (n = 988, 56.6% female, mean age 40.9 (SD 13.1) years). Compared with pre- working from home, mean ratings were decreased for overall physical (2.84, SD 0.087) and mental (2.70, SD 0.93) wellbeing. Overall PA and physical exercise decreased, and overall food intake increased (although this was the same for ‘healthy’ and ‘junk’ food).

In three studies, musculoskeletal problems (especially in the spine) were associated with working from home (Moretti et al. 202041) and starting or increasing telework was significantly associated with pain augmentation (Yoshimoto et al. 202146) which was made worse by decreasing PA, working from a sofa or bedroom and being female (RSPH 202142). Working from home was also associated with a reduction in PA and work comfort, which was further associated with musculoskeletal pain (as a result of lockdown) (Argus et al. 202140). However in one further study, no association between working from home and general body pain was found (Wilke et al. 202144).

Moretti et al. (2020)41 conducted a cross-sectional survey of Italian administrative officers who moved to work online during the COVID-19 pandemic (n = 51, 56.9% female, mean age 46.7 (SD 11.3) years). In relation to health problems, 70.5% of participants reported MSK pain (41.2% back, 23.5% neck, 7.8% shoulder, 7.8% hip, 7.8% knee, 5.9% thigh and 3.9% elbow). Low back pain and neck pain were more severe and interfered more with everyday activities. Neck pain worsened in 50%, improved in 8.3% and was the same in 41.7% of participants, whereas lower back pain worsened in 38.1%, improved in 14.3% and was the same in 47.6% of participants. Home workers without pain reported significantly higher job satisfaction than those with pain.

Yoshimoto et al. (2021)46 conducted an online survey of workers in Japan (n = 1941, 29.5% female, median age 43 (IQR 33, 52)). Starting or increasing telework was significantly associated with pain augmentation in a logistic regression analysis (OR 2.32, 95% CI 1.79 to 3.02), including after adjustment for confounding factors (adjusted OR 2.27, 95% CI 1.68 to 3.06).

Argus et al. (2021)40 conducted an online survey of office workers in Estonia (n = 161, 64.6% female, mean age 38.2 (SD 9.5) years). PA was assessed by the Baecke Physical Activity Questionnaire (BPAQ). Self-reported PA was significantly lower during than before the lockdown, in terms of sport-related PA (mean change in BPAQ −0.52, SD 0.98, 95% CI −0.67 to −0.37, p < 0.001, Cohen’s d = 0.42 (small effect)) and total PA (mean change in BPAI −0.41, SD 1.37, 95% CI −0.62 to −0.19, p < 0.001, Cohen’s d = 0.26 (small effect)), but not leisure-time PA (mean change in BPAQ −0.07, SD 0.59, 95% CI −0.16 to 0.02, p = 0.15, Cohen’s d = 0.11), and work-related PA significantly increased (mean change in BPAQ 0.18, SD 0.54, 95% CI 0.10 to 0.26, p < 0.001, Cohen’s d = 0.50 (medium effect)). There was a significant negative correlation between change in self-reported sports-related PA and change in the numbers of body regions with MSP during the lockdown (r = −0.206, p < 0.01). The number of body regions with MSP onset during the lockdown was also negatively correlated with change in workplace comfort score (r = −0.262, p < 0.001) and change in workplace ergonomics score (r = −0.231, p < 0.01).

Wilke et al. (2021)44 conducted an online survey of people aged 18 plus in 14 countries worldwide with confinement measures limiting movement in public spaces (n = 14,975, 58.1% female, mean age 38 (SD 15) years). Physical wellbeing, assessed using the bodily pain subscale of the SF-36, two items on musculoskeletal pain (6-point scale) and the resulting disability (5-point scale).No associations with physical wellbeing (bodily pain) were found for work mode (p = 0.76).

Mental health and wellbeing

The largest volume of evidence identified consisted of studies conducted during the COVID-19 pandemic which looked at factors which influence the associations between working from home and measures relating to mental health and wellbeing. The factors reported in these studies are summarised in Table 10. Reported factors which influence the associations between working from home and wellbeing or mental health (COVID-19 papers) and Figure 6. Mind map of summary factors which influence the associations between working from home and wellbeing or mental health (COVID-19 papers). A broad range of measures relating to wellbeing were used by study authors including direct measures of wellbeing, and measures of mental health (including negative affect, anxiety, depression, psychological distress), and stress (including perceived stress, perceived self-efficacy, ‘stress, worry and pressure’, burnout, ‘cognitive worsening’ and specific stress including parenting stress and occupational stress). Measures linked to wellbeing included the lifestyle behaviours sleep (sleep quality time sleeping and fatigue), alcohol use and PA (opportunity to exercise, sedentariness, standing and movement, lack of sports facilities open). Increased health concerns were also reported, as were factors linked with social interaction (social isolation, loneliness, minimal contact with others and social support). Satisfaction with working from home (including sense of worthwhile life) and QoL measures were also reported. Lack of choice over whether to WFH along with work autonomy and measures linked to videoconferencing (videoconference fatigue and technostress) were also considered in respect to wellbeing. Further measures linked to the home work environment and wellbeing included feeling in control of time, lack of commute, more time with the family, lower work/family conflict, and spaces shared with others. Openness to new ways of living was also included.

FIGURE 6.

Mind map of summary of factors which influence the associations between working from home and wellbeing or mental health (COVID-19 papers).

Wellbeing no associations

Direct measures of wellbeing were reported in 24 studies, with seven papers reporting no overall association between working from home and wellbeing (Allen et al. 2021,47 Cotterill et al. 2020,50 Di Tecco et al. 2021,35 Gijzen et al. 2020,54 Hoffman et al. 2021,55 Bentham et al. 2021,48 Dunatchik et al. 202153). The potentially relevant variables of having a pet or dependents in the home (Hoffman et al. 2021,55 Bentham et al. 202148) and division of domestic labour (Dunatchik et al. 202153) were also found to not be associated with wellbeing when working from home.

Allen et al. (2021) 47 undertook an online survey (recruited through social media and university courses) in the UK (n = 200, mean age 24.7 (SD 7.2) years, 86% female); 47.0% reported working from home. Psychological wellbeing was assessed using the PWB18 measure. Participants working from home did not significantly differ on wellbeing-related outcome measures. They report that a ‘linear regression model with self-isolation predicting PWB18 scores was significant [F(1,186) = 20.53, p < 0.001]; however, self-isolation (β = −.162, p = 0.004) and UCLA3 (loneliness) scores (β = −.596, p < 0.001) were the only significant predictors’. Cotterill et al. (2020)50 conducted an online survey of water-sector employees in the UK (n == 502, 60.6% male, 84.2% worked from home). One question related to wellbeing: ‘My general wellbeing has improved since lockdown’, and was rated on a Likert scale. More women saw a decrease in wellbeing (39%) than men (32%), although this was not statistically significant, and there were no significant differences between the median wellbeing values for men and women (U = 27 030, z = −1.472, p = 0.141).

Di Tecco et al. (2021)35 conducted a prospective cohort study to investigate the impact of a ‘smart working pilot’ (flexible home working) in Italy (n = 187, 78.6% female, mean age 50.7 (SD 6.8) years). There was no significant change in general health (p = 1.00) or wellbeing (p = 0.247) as a result of the pilot as evaluated by t-tests. Gijzen et al. (2020)54 conducted an online survey in the Netherlands (n = 1519, mean age 53 (SD 16) years (range 18 to 91), 52% female). They aimed to investigate the mental health and wellbeing of adults in the Netherlands 10 weeks after the start of COVID-19. However, although working from home was reported as a positive outcome of the COVID-19 pandemic for 17% (n = 142) participants this was not directly linked to wellbeing.

Hoffman et al. (2021)55 conducted an online survey in the USA with individuals who had experience working from home and from their employer’s office. (n = 454, 231 female, mean age = 41.3 years, SD = 11.5 years). They aimed to ‘Explore perceptions regarding how companion animals factor into the teleworking Experience’. Paired-samples t-tests indicated that neither positive nor negative wellbeing scores differed significantly by workplace location (PAWB: t = 1.17, df = 453, p = 0.24; NAWB: t = −1.74, df = 453, p = 0.08). When analyses were restricted to when participants worked from home, neither positive nor negative wellbeing scores were associated with the presence of dogs or cats in the home.

Bentham et al. (2021)48 completed a survey of people working in a UK mental health service (n = 51, 72.5% female, 29.4% aged 35–44, 23.5% aged 45–54, 15.7% aged 55–64 and 2% aged ≥65 years). Wellbeing was assessing using the Warwick-Edinburgh Mental Well-Being Scale (WEMWBS). An independent-samples Kruskal–Wallis H-test showed no statistically significant differences in wellbeing score based on the proportion of hours worked remotely during the pandemic (χ² (4) = 4.45; p = 0.349). Dunatchik et al. (2021)53 conducted an online survey with a nationally representative sample in the USA (478 partnered parents and 151 single parents). The authors report wellbeing in terms of pressure on working parents, and 66% mothers and 65% fathers reported feeling ‘some’ or ‘a lot’ of pressure regarding children’s home learning during the pandemic. However, no further indication of the impact on wellbeing is presented.

Wellbeing negative associations

However, 11 studies (including a number of grey literature sources) reported a negative association between working at home and wellbeing (Schifano et al. 2021,59 Xiao et al. 2021,45 Ingusci et al. 2021,56 De Sio et al. 2021,51 Wilke et al. 2021,44 Delfino and van der Kolk 2021,52 CIPD 2021,49 Felsted et al. 2020,17 KCL 2021,101 Parry et al. 2021,57 RSPH 202142). A variety of variables were identified which also had a negative effect on the relationship between working at home and wellbeing. Interruptions and distractions as a result of a lack of private space also impacted negatively on wellbeing when working at home (Xiao et al. 202145), as did increased monitoring (by the employer), including an expectation to ‘always be available’, and increased video calling (Delfino and van der Kolk 2021,52 Di Tecco et al. 202135). Higher job demands during the pandemic and work overload also impacted negatively on the work at home / wellbeing relationship (Ingusci et al. 2021,56 De Sio et al. 202151). Further factors which negatively moderated the relationship between working at home and wellbeing included hybrid working (Wilke et al. 202144) along with not feeling ‘sheltered at home’, smoking more cigarettes during the pandemic, and experiencing more psychological distress (De Sio 202151).

CIPD (2021) 49 conducted interviews with 32 senior managers and directors and an online survey with a total sample size of 2133 senior decision-makers in UK organisations, presented as a report and linked web page (grey literature). The most frequently mentioned benefit of working at home was increased wellbeing through avoiding the commute (46% of survey participants), followed by enhanced wellbeing because of greater flexibility of hours (39%).

Delfino and van der Kolk (2021)52 conducted semi-structured interviews with employees from professional service firms in Italy (n = 15, 9 female, 6 male). Management control practices and employees’ responses to these (including wellbeing, and factors affecting wellbeing) were considered ‘to investigate how remote working impacted the use of management control in professional service firms and explore how these changes affected employees’. Negative impacts on wellbeing were mentioned by respondents in terms of monitoring employees’ online/offline status, meaning employees felt they should always be available.

De Sio et al. (2021)51 conducted a web-based cross-sectional survey with Italian teleworkers (n = 348 (60.52%) females and 227 (39.48%) males, median age of 40 years (IQR: 33–49)). Perceived wellbeing was assessed using GHQ-12. Poor wellbeing was associated with having a higher job demand during the pandemic (OR 2.61; 95% CI 1.10 to 6.19), with feeling ‘not sheltered at home’ (OR 8.80; 95% CI 2.60 to 29.75), with smoking more cigarettes during the pandemic (OR 2.47; 95% CI 1.13 to 5.59), and with experiencing psychological distress (OR 8.01; 95% CI 2.57 to 24.97).

Felstead et al. (2020)17 completed three online survey of UK workers (April, May, June 2020). Wellbeing was assessed using GHQ-12 wellbeing questions. Those who worked mainly at home reported greater difficulties in enjoying normal day-to-day activities compared to those not working at home, (48.2/49.3% vs. 38.5%) and more often felt constantly being under strain and unhappy with life (36.0/33.9 vs. 31.2).

Ingusci et al. (2021)56 conducted an online survey of Italians who experienced remote working or working from home (n = 530, 60.4% female, mean age 39.0 (SD 11.2)). Behavioural stress was assessed using a seven-item validated scale. In a structural model with good fit indices (CFI = 0.96, TLI = 0.95, AGFI = 0.91, RMSEA = 0.05 (90% CI 0.05 to 0.06), SRMR = 0.06), behavioural stress was found to be positively related to work overload (β1 = 0.48, p = 0.015) and negatively related to job crafting (β3 = −0.38, p < 0.000), with a significant and negative indirect effect of work overload on behavioural stress through the intervention of job crafting (β^a×b = −0.07, p = 0.029).

KCL (2021)101 conducted an online survey of 254 large employers in the UK. Their report states that organisations who reported improved morale compared to this time six months ago are considerably more positive about the support provided, particularly for parents and carers. This suggests that offering such support can contribute towards improved morale and wellbeing overall.

Parry et al. (2021)57 analysed 1035 survey responses of an UK online worker wellbeing survey. Using the World Health Organisation Five Wellbeing Index (WHO-5) to measure wellbeing, on average respondents scored 47 out of 100, relatively low compared to previous UK and Europe-wide surveys. Key determinants of wellbeing included fewer physical health symptoms, higher levels of satisfaction with work-life balance and identifying as extroverts.

RSPH (2021)42 reported on the mental and physical health impacts of home working during COVID-19. Overall, more people felt working from home was better for their health and wellbeing (45%), compared to around one-third (29%) who thought working from home was worse for their health and wellbeing. Home working was said to be having an impact on people’s mental health, with 67% saying they felt less connected to their colleagues and 56% saying they found it harder to switch off.

Schifano et al. (2021)59 conducted a longitudinal online survey in France, Italy, Germany, Spain and Sweden (n = 9700 observations). Wellbeing was assessed in terms of five variables: life satisfaction and life worthwhile, loneliness, depression, anxiety, and working from home. Working from home was associated with lower wellbeing on all five variables – life satisfaction (coefficient = −0.09, p < 0.01), worthwhile (coefficient = −0.07, p < 0.05), not lonely (coefficient = −0.08, p < 0.05), not depressed coefficient = (−0.09, p < 0.01) and not anxious (coefficient = −0.09, p < 0.01), although not working had a greater negative impact. Switching to working from home reduced anxiety (coefficient = 0.05, p < 0.10) but also reduced the sense of a worthwhile life (coefficient = −0.07, p < 0.05), with no significant impact on other wellbeing variables.

Wilke et al. (2021)44 conducted an online survey in 14 countries worldwide (n = 14,975, 58.1% female, mean age 38 (SD 15) years). Mental wellbeing was assessed using the World Health Organization Well-Being Index (WHO-5). Working outside the home vs. working remotely was associated with clinically relevant reductions in mental wellbeing (OR 1.29, 95% CI 1.16 to 1.44), as was working both outside the home and remotely vs. working remotely (OR 1.35, 95% CI 1.23 to 1.47).

Xiao et al. (2021)45 conducted an online survey in the USA of those who had transitioned to WFH during the COVID-19 pandemic (n = 988, 56.5% female, mean age 40.9 (SD 13.1) years). Physical and mental wellbeing were rated overall, relative to their wellbeing prior to WFH on a Likert scale. Compared with pre-WFH, mean ratings were decreased for overall physical (2.84, SD 0.87) and mental (2.70, SD 0.93) wellbeing. Mental wellbeing was significantly correlated with overall PA (r = 0.36, p < 0.01), physical exercise (r = 0.33, p < 0.01), and distractions while working (r = −0.30, p < 0.01). Mean mental wellbeing was lower for those who adjusted their work hours (2.65, 0.95) than those who did not (2.86, SD 0.87, p < 0.01), those who needed to schedule their work hours around others (2.59, SD 0.95) than those who did not (2.77, SD 0.92, p < 0.01), and those who reported somebody in the same workspace while WFH (2.64, SD 0.95) than those who reported a solitary work environment (2.78, SD 0.90, p = 0.04).

Wellbeing positive associations

Three additional studies reported improved wellbeing for those working at home (Cotterill et al. 2020,50 Xiao et al. 2021,45 PWC 202058). These relationships were moderated by a number of factors, including being an essential worker (amongst employees in the water sector; Cotterill et al. 202091) and having higher levels of overall PA and physical exercise (Xiao et al. 202145).

Xiao et al. (2021),45 described above, reported that improved mental wellbeing (F(38, 351) = 5.306, p < 0.001, R² = 0.371) was predicted by increased physical exercise, increased communication with co-workers, and decreased junk food intake, along with being positively affected by having an infant in the home and negatively affected by increased distractions while working. Cotterill et al. (2020),50 described above, reported that essential workers had the largest improvement in wellbeing (29%). PWC et al. (2020)58 produced a web report to provide insights on the remote working experience of 875 workers in Malta (mean 36–45). The largest percentage of those who viewed the experience as having a positive impact on their wellbeing were those who lived alone and were aged 24–35.

Mental health

Two studies reported a negative association between working at home and a person’s overall mental health (Bevan et al. 2020,89 Kotera 2020102). A number of factors further impacting on the mental health of those working at home included those who were younger workers, looking after elderly relatives (but parents were no different from non-parents in this study), living with parents or renting, new home workers, working more than contracted hours, and reduced contact with their boss (Bevan et al. 2020)89. In one further study a negative association between working at home and experiencing negative affect was described (Janssen et al. 202065).

Bevan et al. (2020)89 in their online survey of UK home workers (n = 500) reported that mental health is poorer for younger workers, those looking after elderly relatives (but parents were no different to non-parents), those living with parents or renting, and those new to home working. Kotera (2020)102 in an opinion piece on new ways of working discusses how more attention needs to be paid to the negative impacts of blurred work–home boundaries, fatigue and increased mental demands.

Janssen et al. (2020)65 conducted an ‘ecological momentary assessment’ (n = 101 (34 adolescents, 67 caregivers), parents: 56.7% female, mean age 48.2; adolescents: 64.7% female, mean age 16.0). Working from home was not related to an increase in parents’ negative affect during the COVID-19 pandemic, as compared with pre-pandemic data.

Anxiety and depression

Anxiety and/or depression and their relationship to work at home wellbeing were considered by nine studies. For anxiety, no factors associated with home working were reported in one study (Wickens et al. 2021117), a positive association was reported in a second study (Smith et al. 2021112), but factors with negative associations for home workers were reported in three further studies (Docka‐Filipek et al. 2021,61 Magnavita et al. 2021105, Burstyn et al. 202191). There were additional negative impacts on wellbeing for females compared to males, and those with financial concerns and/or more dependents (Docka‐Filipek et al. 202161), where people experienced intrusive leadership and overtime work (Magnavita et al. 2021105), and where home workers had concerns about return to work and childcare (particularly for females) (Burstyn et al. 202191).

Five studies on depression reported factors with negative associations between working at home and wellbeing which were moderated by being female and/or having financial concerns (Docka‐Filipek et al. 2021,61 Sato et al. 2021a111), where people experienced intrusive leadership and overtime work (Magnavita et al. 2021105), and where home workers had concerns about return to work and childcare (particularly for females) (Burstyn et al. 202191). Only one study reported a positive association between depression and home working (i.e. home working was association with reduced levels of depression) during the pandemic (Smith et al. 2021112).

Burstyn et al. (2021) 91 conducted an online survey of workers in the USA (n = 911). Anxiety and depression were assessed using the HADS. They report that ‘starting or substantially increasing telecommuting appeared to be associated with increased anxiety and depression in both sexes as well, with the effect more prominent among men’. Men (but not women) who identified as essential workers (RR 1.16, 95% CI 0.96 to 1.40), had one-on-one contact with people at work (RR 1.14, 95% CI 0.98 to 1.34), including known or suspected cases of COVID-19 (RR 1.30, 95% CI 0.97 to 1.74), who were hourly employees (RR 1.24, 95% CI 0.96 to 1.60), and did not have access to disability/sick leave through work (RR 1.22, 95% CI 0.93 to 1.60) were more anxious.

Docka-Filipek et al. (2021)61 conducted an online survey of university faculty working from home at the start of the coronavirus pandemic in the USA (n = 345, 77% female, mean age 42.84 (SD 9.23)). Self-reported mental health (depression and state anxiety) was assessed using a shortened version of the Center of Epidemiologic Studies Depression scale and the state anxiety subscale of the six-item version of the State Trait Anxiety Inventory. Two multiple regression models were run, which accounted for significant variance in depressive symptoms (F(6, 322) = 7.29, p < 0.001) and state anxiety (F(6, 322) = 5.93, p < 0.001). Gender accounted for unique variance in both depression (β = 0.17, p ≤ 0.01) and anxiety (β = 0.17, p ≤ 0.01) risk, after covarying for ethnicity, academic position, teaching load, number of dependents in the home, and financial concerns at separate steps. Higher financial concerns accounted for unique risk of both depression (β = 0.30, p < 0.001) and anxiety (β = 0.26, p < 0.001), and having more dependents accounted for unique risk of anxiety (β = 0.13, p < 0.05).

Magnavita et al. (2021)105 conducted a survey of people working for trade and service-sector companies in Italy who telecommuted part time (n = 905, 36.6% male, mean age 45.93 (SD 11.39)). Common mental health issues were assessed using the Goldberg Anxiety and Depression Scale. They report that anxiety increased with intrusive leadership, workaholism and age, and was higher in females. Depression decreased with off-time work, was higher in females, and increased with workaholism and age.

Smith et al. (2021)112 conducted an online survey with non-healthcare workers in Canada (n = 3305, 61% female, 15.4% aged <34, 23.7% aged 35–44, 29.6% aged 45–54, and 30.3% aged above 55 years). Anxiety and depression were assessed by the Generalised Anxiety Disorder (GAD-2) and the Patient Health Questionnaire (PHQ-2) measures. Among those working remotely, the adjusted proportion of respondents with GAD-2 scores of ≥3 was 35.3% (95% CI 27.1 to 43.5) and the adjusted proportion of respondents with PHQ-2 scores ≥3 was 27.4% (95% CI 20.1 to 34.8), both of which were significantly lower than among site-based workers or those no longer employed.

Wickens et al. (2021)117 conducted an online survey with Canadian workers (n = 1002, 49.7% female, 13.2% aged 18–29, 26.1% aged 30–39, 23.9% aged 40–49, 17.7% aged 50–59, 30.4% aged over 60 years). Depressive symptoms were assessed by a single item from the Center for Epidemiologic Studies Depression Scale. In the regression analysis, after adjusting for demographic variables, working from home was not a significant predictor of depressive symptoms (adjusted OR 1.16, 95% CI 0.75 to 1.77) (the odds of experiencing depressive symptoms were higher among whose with a job at high risk of exposure to COVID-19 and who experienced financial worry due to COVID-19).

Sato et al. (2021a)111 conducted an online survey in Japan (n = 2846, 60% female, mean age 43.0 (SD 12.0) years for females and 50.3 (SD 10.2) years for males). The association of changes in work and life patterns with depressive symptoms was examined in users of the health app CALO mama, which records diet, exercise, mood and quality of sleep and provides feedback to users. Depressive symptoms were assessed by two validated items (‘During the past month, have you often been bothered by feeling down, depressed, or hopeless?’ and ‘During the past month, have you often been bothered by little interest or pleasure in doing things?’). In the logistic regression model, shifting to WFH was negatively associated with depressive symptoms (OR 0.83, 95% CI 0.69 to 0.99).

Other mental health measures

Mental health was also reported as psychological distress. One study reported a positive association with psychological distress (i.e. less distress) for those working at home which was further improved by hybrid working arrangements (alternative office/home working days) (Shockley et al. 202129). However, three studies reported factors with negative associations with working from home, which were moderated by increased childcare (including home schooling) and domestic duties particularly for mothers and lone parents (Clark et al. 2021,60 Xue et al. 202167), and being the only member of couple to adopt their working pattern for childcare (Xue et al. 202167). In addition, increased health concerns were also reported to negatively impact on the mental health of those working at home (Waizenegger et al. 202043)

Clark et al. (2021)60 conducted qualitative interviews in working mothers in Ireland (n = 30). Most participants reported increased levels of psychological distress as a result of the pandemic and resultant changes to the dynamic of work and family life. An additional challenge was managing the psychological welfare of their children during the pandemic, and helping them to cope.

Shockley et al. (2021)29 conducted a longitudinal online survey in the USA of heterosexual married couples where both spouses worked full-time (n = 274 at T1, 133 at T2, mean age at T2: 35.2 (SD 3.4) years for wives and 36.3 (SD 7.0) years for husbands). Psychological distress was assessed by the Kessler et al. (2002) 10-item measure. In the latent class analysis, for health outcomes (psychological distress and sleep quality), those adapting the strategy of ‘alternating days’ fared the best (mean PD score 1.54 and 1.58 for wives and husbands, respectively).

Waizenegger et al. (2020),43 outlined above, report that ‘participants’ mental and physical wellbeing was (equally) affected by a lack of PA, due to sports facilities being closed and minimal contact with others being allowed’.

Xue et al. (2021)67 conducted a longitudinal survey in a nationally representative survey of >100,000 individuals from 40,000 households in the UK (n = 14,150: 8291 women, 5859 men) to describe how men and women divided childcare and housework demands during the height of the first COVID-19 lockdown in the UK. Adapting work patters due to childcare / home schooling was associated with 1.39 (95% CI 0.403 to 2.382) higher GHQ scores in women and 1.16 (95% CI 0.296 to 2.015) higher GHQ scores in men. Being the only member of the couple to adapt a working pattern to accommodate childcare was associated with 1.82 higher GHQ scores (95% CI 0.669 to 2.973) in women and 2.48 higher GHQ scores (95% CI 1.367 to 3.601) in men. Lone mothers who adapted work patters to accommodate childcare / home schooling had on average 3.93 higher GHQ scores (95% CI 1.639 to 6.223; p = 0.001) than lone mothers who did not adapt work patterns. There was no effect of adapting work patterns on GHQ in couple mothers.

Stress

Stress, and its assumed impact on wellbeing, was measured by 13 studies which considered the impact of working from home. Stress was reported as perceived stress, perceived self-efficacy, ‘stress, worry and pressure’, burnout, and as specific stresses including parenting stress and occupational stress. Overall, more papers reported that working from home in the COVID-19 era increased stress than reported reduced stress or no impact on stress from working at home.

No change in perceived stress or perceived self-efficacy was reported with working from home in one survey study, even when differences by professional group (practitioners, managers, executives and teachers) were considered (Mari et al. 2021106). However, factors with negative associations between working at home and stress were reported by five studies (Russo et al. 2021,110 Travers et al. 2020,115 Delfino and van der Kolk 2021,52 Galanti et al. 2021,98 Heiden et al. 2021100). An increase in stress, worry and pressure as a result of working from home was made worse by new and excessive work demands according to Travers et al. (2020). 115 The same relationships were also reported by Delfino and van der Kolk (2021). 52 Negative relationships were also reported in a further three studies with the working from home / stress association mediated by family–work stress, social isolation, and a distracting working environment (Galanti et al. 202198), ‘indistinct organisation’ and fatigue (Heiden et al. 2021100).

Delfino and van der Kolk (2021)52 (qualitative field study) found that employees experienced stress in relation to increased demands and fear of management, which led them to miss breaks to increase their availability, decreased their motivation for their job (and subsequently looked for other work), and had a serious impact on mental health. Galanti et al. (2021)98 conducted an online survey of Italians working from home full time, in public and private organisations (n = 209, 71.3% female, mean age 49.8 (SD 9.4) years (range 25 to 65)). Stress was assessed using four items designed to measure workers’ perception of exhaustion and fatigue due to WFH. Stress was positively correlated with family–work conflict (r = 0.50, p < 0.01), social isolation (r = 0.62, p < 0.01) and distracting working environment (r = 0.36, p < 0.01), and negatively correlated with productivity (r = −0.39, p < 0.01) and work engagement (r = −0.47, p < 0.01).

Heiden et al. (2021)100 conducted an online survey of teaching and research staff at Swedish public universities (n = 392, 63% female, mean age 48.9 (SD 9.9) years). Separate ANOVAs for each outcome variable did not show any significant differences in health (or GHQ subscales), work stress related to individual demands and commitment, and influence at work, or rest, but did show significant differences on fatigue (F = 3.47; p = 0.032) and work stress relating to indistinct organisation and conflicts (F = 4.80; p = 0.009). Post hoc tests revealed that those who teleworked several times per week or more reported more stress relating to indistinct organisation than those who teleworked less than once a month. There were no significant pairwise differences for fatigue.

Mari et al. (2021)106 conducted an online survey in people doing ‘smart work’ (working from home only) (n = 628, 78% female, mean age 42.3 (SD 10.5), age range 21–70). Coping with stress was assessed using the Coping Orientation to Problems Experienced. There were no significant differences between professional groups on the PSS (perceived stress), or on the perceived self-efficacy subscale. For the perceived helplessness subscale, teachers had a higher mean score (11.07, SD 3.90) than managers (9.79, SD 3.81). Russo et al. (2021)110 undertook an international longitudinal survey (mainly UK, USA and countries in Europe) of software professionals working from home during COVID lockdown (n = 192, mean age 36.65 (SD 10.77) years, 20% female). Stress was assessed using a four-item version of the Perceived Stress Scale. At Wave 1, stress negatively affected social contacts, and daily routines predicted stress at α = 0.05. At Wave 2, need for competence and autonomy, stress, quality of social contacts, and quality of sleep uniquely predicted wellbeing at α = 0.05.

Travers et al. (2020)115 conducted an ‘internet-based ethnography’, although little methodology is given (n = 211), of people working at home during the COVID-19 pandemic in the UK. Working from home during lockdown was interpreted as ‘a time of contradictions and transitions’, with new and excessive demands creating worry, stress and pressure, but also opportunities afforded by a lack of commute and spending more time with the family.

Conversely, reduced stress was reported as a positive outcome of home working in four studies (Hayes et al. 2021,99 Molino et al. 2020,107 Delanoeije et al. 2020,94 Rodriguez et al. 2020109), with hybrid working arrangements further reducing stress beyond full-time working at home (Rodriguez et al. 2020109). Ignacio Gimenez-Nadal et al. (2020)64 also reported reduced stress for male teleworkers only.

Delanoeije et al. (2020)94 conducted a quasi-experimental study in Belgium of employees at a large construction and property-development firm allocated teleworking (at least 2 days per week) or control (no home working (n = 78, 75.6% male)). The univariate F tests showed there was a significant interaction effect between time and group for stress (F(1,62) = 4.21, p = 0.04, ηp² = 0.06)’, whereby stress decreased among the teleworking group but not the control group – once commuting time was included as a covariate, there was no group by time interaction effect for stress, suggesting that the decrease in stress among the teleworking group could be accounted for by pre-existing differences in commuting time. For daily stress, ‘the standardized estimate of teleworking day on daily stress (γ = −0.20, p < 0.001) was negative and significant’, supporting the hypothesis that the intervention group would have less daily stress on a teleworking day compared with a non-teleworking day.

Hayes et al. (2021)99 conducted an online survey of people working from home due to COVID-19 restrictions in the USA (n = 326, 52.2% female, 30.7% aged 18–34, 38.3% aged 35–54 and 31.0% aged over 55 years). Stress was assessed by the Perceived Stress Scale (PSS). There was a significantly greater increase in perceived stress score from pre-COVID (retrospectively rated) to the current time among those whose job typically did not provide opportunities to WFH (mean increase 3.9, SD 6.4) than those whose did (mean increase 2.4, SD 5.3) (t(290) = 2.23, p = 0.03). Conversely, those who previously had flexibility to WFH before the pandemic had higher work-related burnout scores at data collection (mean 57.9, SD 21.5) than those without the flexibility to WFH (mean 41.0, SD 21.6) (t(284) = −16.84, p < 0.0001). Although women had lower pre-COVID and during-COVID perceived stress scores than men, the mean increase in stress scores was higher for females (4.2, SD 6.0) than males (2.4, SD 5.8) (t(294) = 2.59, p = 0.01). Women had significantly lower mean work-related burnout scores (43.3, SD 20.8) than men (53.0, SD 24.6) (t(299) = −3.82, p < 0.0002).

Molino et al. (2020)107 conducted an online questionnaire in Italy of home and traditional working (n = 749, 59% female, mean age 38.7 (SD 11.3) years, 63% working from home for a mean 4.7 (SD 1.3) days per week). Technostress creators were measured by the validated 11-item brief Italian technostress creators scale. Behavioural stress was measured by eight items from the Copenhagen Psychological Scale. Significant positive correlations were found between behavioural stress and work–family conflict (r = 0.23), the three techo-stress creators (techno-overload, techno-invasion and techno-complexity; r = 0.22, r = 0.24 and r = 0.23, respectively), and workload (r = 0.19) (all p < 0.01). Work–family conflict was also positively correlated with the three techno-stress creators (r = 0.35, r = 0.48 and r = 0.19, respectively) and workload (r = 0.47) (all p < 0.01). Remote working was positively correlated with techno-overload (r = 0.29), techno-invasion (r = 0.25) and workload (r = 0.13) (all p < 0.01), but not behavioural stress (r = −0.07), work-family conflict (r = 0.03) or techo-complexity (r = 0.01).

Rodriguez et al. (2020)109 conducted an online survey of people living in Spain during the COVID-19 lockdown (n = 1269, 18% male, mean age 38.8 (SD 10.6) years). Stress was assessed by the 14-item Perceived Stress Scale. Working situation (during COVID-19 lockdown confinement) was related to stress response (F(4,918) = 4.914; p < 0.01; ηp² = 0.020) and control of stress (F(4,928) = 4.017; p < 0.01; ηp² = 0.016).

Ignacio Gimenez-Nadal et al. (2020)64 conducted a face-to-face survey of employee workers in the USA (n = 5401, 47.9% female, mean age 43.8 (SD 10.7) for males and 43.9 (11.0) years for females). This consisted of recording a daily diary where pain, happiness, sadness, fatigue and stress were each rated 0–6. They report that among males, teleworkers reported lower levels of sadness, stress and tiredness compared with commuters. Among females, teleworkers had significantly higher happiness levels than commuters. There were no other significant differences.

In terms of specific stressors, parenting stress (made worse by lower PA) was negatively associated with working from home (Limbers et al. 202066). Evans et al. (2021)96 reported that those who scored highly for the personality measures of extroversion and conscientiousness were more likely to experience burnout while working from home as the pandemic progressed.

Limbers et al. (2020)66 conducted an online survey in the USA with females with at least one child who would normally work outside the home but were working from home during the COVID-19 pandemic (n = 200, mean age 33.5 (SD 6.25) years). Parenting stress was assessed using the Parental Stress Scale. Greater levels of parenting stress were associated with lower physical health QoL (r = −0.42, p < 0.001), lower psychological QoL (r = −0.28, p < 0.001), lower social relationships QoL (r = −0.21, p < 0.01), and lower environment QoL (r = −0.19, p < 0.01). Evans et al. (2021)96 conducted an online survey in UK remote workers (n = 974, 61% female, mean proportion of time WFH 91%). Burnout was assessed over the preceding month using five items. At the first wave, those scoring high on extroversion and conscientiousness were less likely to experience burnout, whereas those high in extraversion reported higher levels of burnout over time.

Satisfaction

Satisfaction with working from home (including sense of worthwhile life) and QoL measures were also reported. Measures of life satisfaction, along with a sense of having a worthwhile life, were reduced with working from home in four studies (Hubbard et al. 2021,63 Toscano et al. 2020,114 Taser et al. 2022113). Satisfaction with working from home was lower for females (Hubbard et al. 202163), those with caring responsibility (Hubbard et al. 202163) and those who had a lack of suitable space to work in the home (Hubbard et al. 202163). Low enjoyment of activities as a result of working from home was also reported (Felsted et al. 202017). Openness to new ways of living was positively associated with a positive view of working from home (Boncori 202090).

Boncori (2020)90 wrote an auto-ethnography on the experiences of living and working during the COVID-19 pandemic (early stages). She wrote of how openness to new ways of living was positively associated with a positive view of working from home. Felstead et al. (2020)17 (online survey) found that those who worked mainly at home – always or often – reported greater difficulties in enjoying normal day-to-day activities compared to those not working at home (48.2/49.3% vs. 38.5%) and more often felt constantly being under strain and unhappy with life (36.0/33.9 vs. 31.2).

Hubbard et al. (2021)63 conducted a cross-sectional survey (online) including 501 adults of working age currently working in the UK. Ratings of satisfaction with working from home were sought. Women reported less satisfaction than men (chi-square 7.011, df = 3, p = 0.071), as did people with children (chi-square 7.299, df = 3, p = 0.063) – especially young children aged 0–4 years (chi-square 8.01, df = 3, p = 0.046). A significant predictor of dissatisfaction with home working was caring for a responsible adult (chi-square = 7.837, df = 3, p = 0.049). No other predictors were listed.

Toscano et al. (2020)114 conducted an online survey of Italian employees working exclusively from home (n = 265, 63% female, 42% aged 26–35, 21% aged 36–45, 17% aged 46–55, 11% aged <25 and 8% aged over 56 years). Perceived remote working productivity, remote job satisfaction, concern about COVID-19, and experience with remote work were also assessed. Social isolation was significantly correlated with stress (0.50, p < 0.01), perceived remote work productivity (−0.43, p < 0.01), remote work satisfaction (−0.50, p < 0.01) and COVID-19 concern (0.32, p < 0.01). Stress was significantly correlated with perceived remote work productivity (−0.35, p < 0.01), remote work satisfaction (−0.54, p < 0.01) and COVID-19 concern (0.16, p < 0.05).

Taser et al. (2022)113 conducted an online survey of Turkish financial services sector employees, working from home during the pandemic (n = 202, 51.5% female, 2.4% aged 18–25, 25.7% aged 26–35, 42.2% aged 36–45, and 29.7% aged over 46 years). Technostress was assessed using the Tarafdar scale. Those who had a good remote e-working experience tended to have lower levels of technostress (b = −0.17, SD = 0.06, p < 0.01). Those experiencing technostress were likely to feel lonely (b = 0.23, SD = 0.06, p < 0.001).Technostress and loneliness mediated the relationship between remote e-working and flow at work.

Lifestyle behaviours

Lifestyle behaviours including sleep (and fatigue), alcohol use and PA (or sedentariness) were considered in seven studies. Sleep (sleep quality, time sleeping and fatigue) increased as a result of working from home in two studies (Shockley et al. 2021,29 Hallman et al. 2021118); however, it was not entirely clear whether an increase in sleep duration (rather than sleep quality) was considered to be positive. However, the impact of working from home on sleep was described as most positive for hybrid workers (alternative days at home/office) (Shockley et al. 202129). There was a negative association between levels of fatigue and working from home (Heiden et al. 2021100) and fatigue increasing whilst working from home reported in a further two studies (Kotera 2020,102 Parry et al. 202157). Ignacio Gimenez-Nadal et al (2020)64 reported reduced fatigue for male teleworkers only.

Hallman et al. (2021)118 conducted a cross-sectional online survey, with diary and accelerometer data, comparing activity between working in the office and working from home in Sweden (n = 27, 81.5% female, mean age 43.4 (SD 9.9)). Accelerometer-assessed PA (proxy for physical health), standing, sedentariness and sleep were assessed. Sedentariness, standing and movement did not differ significantly between working from home (WFH) and working at the office (WAO). Time spent sleeping (relative to time spent awake) was significantly greater on working-from-home days than for days working at the office. Days working from home were associated with more time spent sleeping relative to awake, and the effect size was large (F = 7.4; p = 0.01; ηp² = 0.22). The increase (34 min) in sleep time during WFH occurred at the expense of a reduction in work and leisure time by 26 min and 7 min, respectively.

Five studies have been discussed previously: Heiden et al. (2021)100 (online survey) found home workers did show significant differences on fatigue (F = 3.47; p = 0.032). Kotera 2020102 (opinion piece) reported that new ways of working increase blurred work–home boundaries, fatigue and mental demands. Parry et al. (2021)57 (online survey) found higher levels of fatigue in home workers. Ignacio Gimenez-Nadal et al. (2020)64 (face-to-face survey) report that among males, teleworkers reported lower levels of tiredness compared with commuters. Shockley et al. (2021)29 (longitudinal online survey), in their latent class analysis, found that for health outcomes (psychological distress and sleep quality), those adapting the strategy of ‘alternating days’ fared the best (mean PD score 1.54 and 1.58 for wives and husbands, respectively).

The impact of PA levels on wellbeing whilst working from home were considered in a couple of different ways. Levels of sedentariness, standing and movement were not associated with wellbeing in one study (Hallman et al. 2021118), which included validating self-reported outcomes with accelerometer data. However, having more opportunity to exercise whilst working at home was positively associated with wellbeing (Travers et al. 2020115) and the lack of sports facilities open during lockdown was negatively associated with wellbeing due to reduced PA levels (Waizenegger et al. 202043). Greater alcohol use was associated with reduced wellbeing for those working at home in one study (Alpers et al. 202187).

Alpers et al. (2021)87 conducted an online survey in a random sample of adult residents in Norway (n = 25,708, 56% female, 13% aged 18–29, 16% aged 30–39, 18% aged 40–49, 21% aged 50–59, 18% aged 60–69, 14% aged 70 plus). Alcohol consumption was assessed through the AUDIT-C. Self-assessed increased alcohol consumption during the lockdown period was more frequently reported by people working or studying from home (OR 1.4, 95% CI 1.3 to 16) (as well as those reporting economic worries and in quarantine).

Hallman et al. (2021)118 (previously described: cross-sectional online survey) found that sedentariness, standing and movement did not differ significantly between working from home (WFH) and working at the office (WAO). Time spent sleeping (relative to time spent awake) was significantly greater on working from home days than for days working at the office. Travers et al. (2020)115 and Waizenegger et al. (2020)43 have also been previously described.

Social interaction

Social interactions were reported as social isolation, loneliness, minimal contact with others and social support. Social isolation was reported as more likely for those working at home (Lal et al. 2021104) and the relationship was mediated by technology and silence (Gao and Sai 202062), perceived remote work productivity and remote work satisfaction (Toscano et al. 2020114). Loneliness also increased during working from home, as was noted in four studies (Wood et al. 2021,119 Taser et al. 2022,113 RSPH 2021,42 University of Exeter 2020116), and minimal contact with others was made worse by reduced PA (Waizenegger et al. 202043). Conversely, the participants of two studies reported that their social support increased as a result of working at home (Wood et al. 2021119, Russo et al. 2021110).

Lal et al. (2021)104 conducted an exploratory, interpretive qualitative diary-keeping study of people who had recently transitioned to WFH during the pandemic mostly in the UK (n = 48% female, 69% had previously worked from home). Maintaining interaction while working remotely was also an issue, with a lack of face-to-face interaction leading to worries about colleagues: some people missed the small daily social interactions that they usually had at work. Gao and Sai (2020)62 completed an auto-ethnography (personal reflections of three researchers) in the UK. No details of the analysis were given. Both women experienced social isolation as a result of being physically distanced from their workplace and colleagues, even if working alone was previously sought. This was also explored in the context of possible regret for choosing to live alone (e.g. rather than starting a family), as one thing that added to the sense of isolation was not being able to see other people or to hug them.

Wood et al. (2021)119 reported a 4-week diary study over two time periods among UK university staff (n = 784 (20% response)). The Warwick–Edinburgh Mental Well-being Scale (WEMWBS) was used to assess the impact of home working on home worker wellbeing. Of the predictors tested at the between-person level, loneliness was associated (negatively) with all wellbeing measures for both phases.

The University of Exeter (2020)116 report a weekly survey during COVID-19 lockdown on their webpage of university employees working from home (n = 85). They state that the loneliness of working in a home environment and increased demands to juggle work and domestic responsibilities caused a decline in employee wellbeing. Nearly one in five (17%) remote workers reported feeling lonely,

In previously described studies: in Toscano et al. (2020)114 (online survey) social isolation was significantly correlated with stress (0.50, p < 0.01); Taser et al. (2022)113 (online survey) reported that technostress and loneliness mediated the relationship between remote e-working and flow at work; in RSPH (2021)42 (online report) nearly half (48%) of people who worked from a sofa or bedroom said they had developed musculoskeletal problems and nearly two-thirds (59%) said they felt more isolated from their colleagues; Waizenegger et al. (2020)43 (qualitative) reported that ‘participants’ mental and physical wellbeing was (equally) affected by minimal contact with others being allowed’; and Russo et al. (2021)110 (longitudinal survey) reported stress negatively affected social contacts.

Home environment

Measures linked to the home-work environment and wellbeing included feeling in control of time, lack of commute, more time with the family, lower work/family conflict, and spaces shared with others. The lack of a daily commute was positively associated with wellbeing for home workers (Travers et al. 2020115). A feeling of being in control of time was also positively associated with wellbeing,115 as was more time with the family (Travers et al. 2020115) and lower work/family conflict (Fukumura et al. 202197). However, having to work in home spaces shared with others was negatively associated with wellbeing (Waizenegger et al. 202043, Fukumura et al. 202197).

Fukumura et al. (2020)97 conducted an online survey of individuals who transitioned to WFH during the pandemic in the USA (n = 988, mean age 41.7 years (SD 12.9), 65.4% female). Physical and mental wellbeing were assessed using ‘Likert-type categorical response questions’. Some people preferred the spatial arrangements at home, for instance those with internal offices in the workplace, and no natural light, where they could work at home in a more comfortable space and take breaks in the garden. However this varied between people, with some describing the home environment less conducive to work, due to the presence of others including children, a lack of privacy and a lack of appropriate technology. Others reported wellbeing benefits in not having to speak to colleagues that they didn’t want to speak to, or be distracted by other people’s conversations in the office.

In previously described studies: Travers et al. (2020)115 (‘internet-based ethnography’) reported that the availability or lack of availability of suitable work space at home also impacted on people’s wellbeing. Working in unsuitable spaces (e.g. landing, blocking fridge door) and competing for space with other family members (e.g. children, pets) could cause problems, but people also expressed warmth for their children and pets; and Waizenegger et al. (2020)43 (qualitative) discussed how participants often worked in a space shared with other occupants of the household, and this caused distraction and difficulty in focusing on work tasks.

Work-related outcomes

Lack of choice over whether to WFH was explicitly stated as a negatively associated with wellbeing in two studies (Waizenegger et al. 202043, Wood et al. 2021119). However, this factor is implicit in all of the studies conducted during the COVID-19 pandemic and as such, although not frequently considered and reported, is likely to be one of the key factors impacting on wellbeing of those working at home. Along with this, increased work autonomy was positively associated with wellbeing for those working at home (Wood et al. 2021119).

Waizenegger et al. (2020)43 (qualitative) note that findings should be interpreted in the context that no-one had a choice about working from home, it was a requirement, and that most of the company’s workforce needed to WFH. Wood et al. (2021)119 (qualitative diary study) state that factors pertaining to the enforced nature of home working and the COVID-19 factors pertaining to increases in deaths and the interaction effect of this with age had some bearing on wellbeing.

Measures linked to videoconferencing including video conference fatigue, further negatively impacted by having camera on, and low sense of belonging as a result of remote interactions (Bennett et al. 2021)88 were all negatively associated with wellbeing for home workers.

Bennett et al. (2021)88 conducted an online survey with closed (quantitative) and open-ended (qualitative) questions with individuals working in a range of industries in the USA (n = 55, 58% male, 73% White, mean age 33.6 years (SD 9.1)). Videoconference fatigue was measured using the fatigue item from the Profile of Mood States scale, scored on a 6-point scale. In qualitative data, participants reported being particularly fatigued by multiple (including consecutive) videoconference meetings. Muting the microphone (γ = −0.09, p = 0.02) and perceptions of group belongingness (γ = −0.21, p = 0.003) were negatively related to fatigue (i.e. were associated with lower fatigue), whereas turning the webcam off, attention during the meeting, and videoconference meeting duration were not significantly related to post-meeting fatigue. The authors tested the interaction between muting and perceptions of belongingness (which were significantly negatively correlated with each other; −0.45, p < 0.05) in a multilevel regression, and the interaction was significant: ‘mute levels do not impact fatigue at high levels of group belongingness, indicating the importance of group belongingness to reduce videoconference fatigue’.

Pre-COVID-19 studies

Qualitative data from pre-COVID working from home research revealed an intricate interplay of benefits and detriments (see Figure 7). Participants described positive impacts on their emotions, which stemmed from being more satisfied with their jobs (or even making a job that they didn’t enjoy bearable; Tietze 201133), attaining a better balance between work and other aspects of their lives, and having the mental space for creativity in their jobs, without the distractions of a workplace (Charalampous 2021,12 Daniel et al. 201816). In some cases, working from home brought more autonomy and control in terms of not just the time and place of work but also workloads, and this sense of control helped people to manage stress by fitting work in flexibly around other commitments and interests, and keeping workloads manageable (Daniel et al. 2018,92 Hislop 2015,21 Mann et al. 2003,25 Tietze 201133). Avoiding negative aspects of the workplace, such as negative interactions like ‘backbiting’, ‘bitching’, ‘bickering’ and ‘gossiping’ (and the resultant stress experienced), contributed to home workers’ wellbeing (Collins et al. 2016,15 Tietze 201133).

However, the flipside to this was that people working from home experienced isolation, frustration and stressors relating to the job role of specific WFH occupations. People working from home experienced isolation and loneliness as a result of the lack of social interaction, including not having day-to-day contact with co-workers, not getting to know new office staff, and no-one to help with problem-solving and ‘bounce ideas off’ (Charalampous 2021,12 Collins et al. 2016,15 Daniel et al. 2018,92 Hislop 2015,21 Mann et al. 2003,25 Stitou 2018,31 Vitterso 2003,10). This was exacerbated by extraversion (Koehne et al. 201223), working long and irregular hours (Daniel et al. 201892) and a longer duration of and/or more time spent working from home (Collins et al. 2016,15 Daniel et al. 2018,92 Vitterso 200310). This isolation could result in poor mental health (Stitou 201831) and in a feeling of insecurity and lack of confidence (Mann et al. 200325), and was even felt when the home-worker was working in home-based childcare and had daily contact with young children, but not other adults (Stitou 201831). People would counteract this isolation by finding ways to maintain communication and/or social contact, whether with their colleagues or in the community around their home, and some people reported that the lack of a commute gave them more time to spend with family and friends (Grant et al. 2013,18 Koehne et al. 2012,23 Tietze 2011,33 Vitterso 200310).

Another detriment to working from home was the feelings of irritation and frustration due to work factors outside the home-worker’s control that impacted on their ability to work. These included frustration with technological issues, not being able to get hold of colleagues when needed, and the intrusion of family members into work time (Charalampous 2021,12 Mann et al. 200325). The nature of the types of work that are amenable to a home-based setting could also cause stress and poor mental and physical wellbeing. For instance, home-based childcare work was perceived as mentally and emotionally draining, due to having to maintain concentration for long periods of time, deal with children with behavioural conditions, maintain relationships with parents, and ensure that regulations were being adhered to in case an inspector called unannounced, all while being (self-)employed precariously with no or little benefits such as sick pay, healthcare or pensions (Stitou 201831). In some types of office work where only a proportion of the workforce worked from home, staff were worried that they may be called back into the office if they were perceived as not performing to a certain standard, which had a negative emotional impact (Collins et al. 2016,15 Tietze 201133). Organisational culture could also impact on the wellbeing of people working from home. In some cases there was a possibility that working from home could impact on career progression and development due to missing out on opportunities and training, and thus being part of an organisation that embraced working from home was seen as being important by workers (Charalampous 202112) (Figure 7).

FIGURE 7.

Thematic structure tree diagram for pre-COVID studies.

COVID-19 studies

A wide variety of issues were reported through qualitative data analysis in relation to working from home during the COVID-19 pandemic. Again, there was a balance of benefits and detriments to working from home during the pandemic, although in this case there seemed to be more detriments (see Figure 8 and Figure 9). This could be due to the emergency situation, where people felt less control overall in their work circumstances, and this was reflected in the difficulty of finding appropriate working space in the home (Fukumura et al. 2020,97 Travers et al. 2020115), particularly having to share the space with other members of the household, including partners, housemates and children (Clark 2021,60 Fukumura et al. 2020,97 Travers et al. 2020,115 Waizenegger et al. 202043). Home workers spoke about the forced situation, and how it blurred boundaries between work and home life (Boncori 2020,90 Clark 202160), with new demands that included having to get used to an unusual situation, excessive work demands (including an expectation from employers of business as usual) and family demands, overlaying a backdrop of general worry and anxiety relating to the pandemic (Fukumura et al. 2020,97 Travers et al. 2020115). The change in routine was something that some people found a struggle; others adapted by introducing new routines (e.g. exercise regimes), but described these as unsustainable over the longer term (Travers et al. 2020115). As with the pre-OVID studies, isolation was experienced while working from home, but this was potentially more intense than pre-COVID due to the duration of homeworking and the lack of possibility for other forms of social interaction (Clark 2021,60 Gao and Sai 2020,62 Lal et al. 2021104).

With all interaction being online, videoconference fatigue was reported as a phenomenon, with symptoms including anxiety and tiredness (Boncori 2020,90 Lal et al. 2021,104 Waizenegger et al. 202043), as well as musculoskeletal pain from having to sit on video calls for long periods (Boncori 202090). There was also an element of occupational work invading the home space (Boncori 202090). Having to try and translate work online was a source of stress, and this was reported among those who had to try and manage the change among a number of employees (Boncori 202090). Additional pandemic-related concerns were silence and awkwardness in online interactions (Gao and Sai 202062) and concern and worry about colleagues and how they were coping, in the absence of visual cues (Lal et al. 2021104). Negative emotional responses such as ‘brain fog’, anxiety and fear were also reported (Travers et al. 2020115).

One thing that seemed more prevalent among the COVID qualitative data than among the pre-COVID data was increased monitoring of staff by managers, to the point where it became excessive and intrusive; people felt under pressure to be available and responsive at all times, and were under almost constant contact and scrutiny in some cases, which caused stress and fear (Delfino and van der Kolk 2021,52 Fukumura et al. 2020,97 Lal et al. 2021104). There was also an element of occupational work invading the home space (Boncori 202090). As with the pre-COVID studies, isolation was experienced while working from home, all the more so because of the duration of homeworking and the lack of possibility for other forms of social interaction (Clark 2021,60, Gao and Sai 2020,62 Lal et al. 2021104). The change in routine was something that some people found a struggle; others adapted by introducing new routines (e.g. exercise regimes), but described these as unsustainable over the longer term (Travers et al. 2020115).

Benefits included improved wellbeing, including through avoiding commuting and some of the negative aspects of office life (Fukumura et al. 202097), increased flexibility (Travers et al. 2020115), and, for some people (e.g. those with internal offices in the workplace), more optimal working conditions, with natural light and the possibility of spending breaks in the garden (Fukumura et al. 202097). To counteract isolation, some people created social interactions with friends, relatives and colleagues, for instance through ‘virtual coffees’, with some employers deliberately implementing these (Boncori 2020,90 Travers et al. 2020,115 Waizenegger et al. 202043). Others appreciated spending more time with family (Lal et al. 2021104). A beneficial side effect of the pandemic among those who were already working from home is that, in some cases, bonding between the whole team improved, and existing remote workers were included in social events for the first time (which would previously have been arranged as ‘in-person’) (Waizenegger et al. 202043).

Some inequalities were highlighted, for instance those with smaller or more crowded living space found working from home more stressful, particularly as pandemic conditions meant the space became crowded (Fukumura et al. 202097). The need to juggle work with childcare/home-schooling seemed to fall primarily onto women (Clark 202160), highlighting a potential gender inequality in working from home during the pandemic. This was exacerbated by a lack of consideration of childcare from male partners’ workplaces (Clark 202160).

Certain elements of organisational culture were found to impact negatively on wellbeing, for instance treatment of fathers vs. mothers, and micro-management. In considering potential childcare responsibilities of female but not of male workers, organisations inadvertently perpetuated gender inequality, which could disproportionately negatively impact women in terms of high stress levels and depression experienced by mothers who have to work at home (Clark 202160). Increased monitoring of staff by managers seemed more of a salient concern among the COVID qualitative data than among the pre-COVID data, to the point where it became excessive and intrusive. This often took the form of increased messages or phone calls from a supervisor and increased numbers of video calls (from the usual number of face-to-face meetings prior to the pandemic) (Delfino and van der Kolk 2021,52 Fukumura et al. 2020,97 Lal et al. 2021104). Staff felt under pressure to be available and responsive at all times, and were under almost constant contact and scrutiny in some cases, which caused stress and fear (Delfino and van der Kolk 2021,52 Fukumura et al. 2020,97 Lal et al. 2021104). This seemed to be born out of a lack of trust (Delfino and van der Kolk 202152), perhaps because of a lack of control over the work situation.

COVID-19 studies

Seven COVID-19 studies reported on variable outcomes by gender, caring responsibilities, age, ethnicity, education and income.

Gender

Five studies reported negative impacts of working at home for women compared to men during the COVID-19 pandemic. Hubbard et al. (2021)63 said that women reported less WFH satisfaction than men during the COVID-19 pandemic. Further, Docka-Filipek et al. (2021)61 found that gender accounted for unique variance in both depression and anxiety (with rates for females higher in both cases). Stress and ‘non-healthy days’ increased in women but not men (Ray et al. 202127) and women were more likely to feel isolated and report developing musculoskeletal problems (RSPH 202142). Cotterill et al. (2020)50 reported that more women perceived that their wellbeing decreased in lockdown than men but found no difference in median wellbeing values. However, one study reported positive outcomes in women compared to men. Sato et al. (2021b38) found a greater increased intake of vegetables, fruits and dairy products and decrease in alcohol intake among women than men. Finally, Ignacio Gimenez-Nadal et al. (2020)64 reported that, compared with commuters, male teleworkers (but not females) reported lower levels of sadness, stress and tiredness and female teleworkers (but not males) reported higher happiness levels.

Burstyn et al. (2021)91 found that anxiety and depression were negatively linked to concerns about return to work, childcare, lack of sick leave, and loss/reduction in work. They reported that these ‘patterns differed by sex’ but the direction of differences was not clear. Finally Weitzer et al. (2021)39 reported that men were more likely to report no changes in QoL than women.

Caring responsibilities

Three studies looked at caring for children. Chung et al. (2020)93 reported that increased workload and conflict between work and family negatively impacted parents’ mental wellbeing, especially for mothers: ‘almost half’ of all mothers felt rushed and pressed for time, more than half of the time during the lockdown, and 46% of mothers felt nervous and stressed more than half of the time. A study by the University of Exeter (2020)116 also reported that increased demands to juggle work and domestic responsibilities (including childcare) had a negative effect on wellbeing. Having more dependents also accounted for ‘unique risk for anxiety’ (Docka-Filipek et al. 2021)61. Two further studies considered caring for adults. Hubbard et al. 202163 reported that a significant predictor of dissatisfaction with home working was having caring responsibilities for an adult. Mental health was found to be poorer among those who were looking after elderly relatives (but parents were no different to non-parents) (Bevan et al. 202089).

Age

Two studies considered age as a variable factor and reported that older participants were most likely to report no changes in QoL from working at home (Weitzer et al. 202139), and also that mental health was generally poorer for younger home workers compared to older age groups (Bevan 202089).

Ethnicity

Only one study considered the potential impact of ethnicity and found that reports of stress and ‘non-healthy days’ increased in ‘non-whites’ (Ray et al. 202127).

Education and income

The same two studies reported that those who were ‘university educated’ were most likely to report no changes in QoL (Weitzer et al. 202139), and that stress and non-healthy days increased in those of lower income and lower health status (Ray et al. 202127) (Figures 8 and 9).

FIGURE 8.

Thematic structure tree diagram for COVID-era studies – Benefits theme.

FIGURE 9.

Thematic structure tree diagram for COVID-era studies – Detriments theme.

Wellbeing measures

A notable omission from the evidence is that the concept of enforced working from home and having ‘no choice’ was reported in only one paper prior to the pandemic and two papers reporting on working from home as a result of COVID-19 and the associated lockdown measures. However, the concept of lack of choice around working from home was implicit in much of the literature published during COVID-19 – even though it was not directly measured.

Prior to COVID-19 one study considered the difference of working from home on a voluntary basis compared to where it was enforced. Involuntary working from home was associated with greater work-to-family conflict, stress, burnout, turnover intentions, and lower job satisfaction whereas voluntary WFH was protective, and associated with greater job satisfaction, lower turnover intentions, and less stress (Kanduk et al. 2019). Two COVID-19 studies reported having no choice but to WFH as negatively impacting on wellbeing (Waizenegger et al. 2020,43 Wood et al. 2021119). The concept of choice was also implicit in many of the other COVID-19 papers, although it was not directly measured. It seems that allowing employees some flexibility over whether they choose to work at home in the future, and beyond the COVID-19 pandemic, will be a key factor in determining whether the home working experience is positive or negative, particularly for mental health and wellbeing-related outcomes.

Home environment

In terms of the home working environment, for factors which were reported both before and during the COVID-19 pandemic, there was no change of direction for any of the measures considered (with the exception of one study). Positive factors reported prior to COVID-19 which were also reported as being positively associated with wellbeing during the pandemic included having control over the workspace and the times worked, not having to commute (and the associated reduction in stress) and experiencing fewer distractions and disturbances whilst working at home, usually due to fewer people in the household (12 studies in total). The association which remained almost entirely negative with respect to wellbeing was family conflict as a result of home working and interference with family life (although one COVID-19 study did report a positive association with family conflict). No other measures relating to the home working environment were considered by studies conducted before or during the pandemic.

Job-related outcomes

Job-related factors which remained positively associated with better wellbeing both before and during the pandemic included hybrid working arrangements (three studies), having high autonomy over work (three studies) and receiving support from the employer to work at home (four studies). High work stress (four studies), high workload (six studies), work demands (four studies), surveillance (including an expectation to always be available) (six studies), low work perceptions (five studies) and problems associated with technology (seven studies) were consistently associated with poorer wellbeing-related outcomes for home workers. There were no consistent patterns in relation to professional group or type of employment.

Individual outcomes

There were consistently positive associations for better wellbeing both before and during the pandemic in terms of having good social support (10 studies), and a personality type which meant an individual had high ‘coping’ (two studies) and ‘openness to new experience’ (two studies). Having financial worries was unsurprisingly consistently associated with poorer WFH wellbeing (two studies). For depression, associations with health outcomes were mostly negative for studies conducted during the pandemic (resulting in poorer health) (seven studies), but positive before COVID-19 (resulting in better health) – although this was only reported by one study prior to the pandemic. All other outcomes were mixed in terms of the direction of association with wellbeing; or were not reported by studies conducted both before and during the pandemic.

Therefore, perhaps surprisingly there were no clear patterns of wellbeing measures which changed from positive to negative association (or vice versa) during the pandemic. This is of course determined by what authors chose to measure and report, and the paucity of evidence on wellbeing measures prior to the pandemic, so should in no way be taken to suggest that pandemic home working did not have an effect on wellbeing overall. Further it is impossible to separate out the effects of COVID-19 lockdown and uncertainties on wellbeing from the direct impacts of home working during this time on wellbeing, particularly for studies conducted during the early stages of the pandemic.

Physical and overall health measures

In terms of physical and overall health measures, the significantly smaller number of studies measuring these types of association both before and during the COVID-19 pandemic made it even more challenging for any potential patterns to be identified. However, there is some indication that the association between working at home and PA measures became more negative during COVID-19 pandemic, with five studies reporting reduced PA (compared to a more mixed picture before the pandemic of two studies reporting factors with positive associations, one negative, and one reporting no effect). This is unsurprising given the lockdown measures in which home working was implemented, during which time sports facilities, leisure centres and gyms were closed.

All other measures did not appear to be affected to any measurable extent by the COVID-19 pandemic (most likely due to the low number of studies identified). Better QoL and life satisfaction remained positively associated with working from home (in one COVID-19 study and two from before the pandemic); studies reporting on general/physical health and burnout all reported factors with negative associations with working from home (two studies from before COVID-19, including one of enforced home working, and three COVID-29 studies). Studies on pain-related outcomes during COVID-19 (four studies) reported factors with negative (less pain) or no association (with no pre COVID-19 studies identified). Studies measuring diet (four COVID-19 studies) showed factors with mixed positive and negative associations, as did studies on substance use, which were mostly positive (one measuring tobacco use during COVID-19, two measuring alcohol use; one before and one during COVID-19), with one final COVID-19 study reporting no association between drug use and working from home.

Changing preferences

Even prior to the COVID-19 pandemic, more people were choosing home working for some or all of their working hours. Between 1999 and 2019 working mainly from home increased by 80% to reach 5.3% of workers (CIPD 2020). 73 However, prior to COVID-19 the majority of home workers did so only occasionally. A key determinant of working from home is age, with older workers more likely to be in the ‘mainly work from home’ category and the ageing workforce a reason for the increase in the number of people working from home (CIPD 2020). 73 Home working also varies by level of skill and experience and by industry, with ‘managers, directors and other senior officials and people in professional occupations’ most commonly working from home, and ‘process, plant and machine operatives and those in elementary occupations’ least likely to do so (CIPD 2020). 73 Flexible work and multi-career approaches have become more mainstream in recent years, with the UK Government now offering guidance that all employees have the right to request flexible working (working in a way of working that suits an employee’s needs, for example having flexible start and finish times, or working from home), with employers having to provide a ‘good business reason’ for refusing an application for flexible working (GOV.UK 2021120).

Technology

A key driver enabling working from home is technological development, with the transition to digital tasks and communication having a profound effect on the number and types of jobs which can be done from home. For most jobs the use of a computer is essential for at least some aspects of the work, and the huge increase in household internet access over the last two decades has enabled many more people to WFH (CIPD 202073).

Commuting

Increased commuting time is another driver, with people who occasionally worked from home prior to COVID-19 having a nine-minute longer journey time than those who do not (CIPD 202073). However, it is not clear whether longer commutes were leading people to WFH, or the ability to WFH was facilitating longer commutes for those with hybrid working arrangements.

COVID-19

Clearly the impact of the pandemic has had a significant effect on increasing the number of people working at home (at least for the present time). As a result, 35.9% of employees did some work at home in 2020, an increase of 9.4 percentage points compared with 2019 (ONS 202176).

Environmental concerns

With the COVID-19 pandemic lockdown causing a measurable decrease in carbon dioxide emissions as a result, in part, of decreased road and air-traffic journeys (including journeys made to and from places of work) (Sikarwar et al. 2021121) the awareness of the impact of commuting and travelling for work on contributing to the global climate emergency has been raised.

The results of our systematic review allowed us to build on the a priori model3 in order to develop an evidence-informed model and to add the huge variety of additional factors reported in the literature which are associated with the pathway between working at home and health-related outcomes. As a result of this analysis we have separated the factors into those related to the home working environment, those related to the job and employer behaviours, and those relating to the individual. Also implicit in the literature identified is the concept that having choice over whether to work at home or not is key to determining whether the experience is likely to be positive or negative for the individual. This extends to the notion that for many, a hybrid working situation (some time at home and some time in the office or usual place of work) can be of greatest benefit. Our final evidence-based model (see Figure 10. Final evidence-informed model of the factors which influence the associations between working from home and health outcomes) therefore draws together the result of our systematic review (and previous mapping review) along with our PPI and stakeholder consultation exercises to provide an overall summary of the reported factors which influence the associations between working from home and health (and related) outcomes.

FIGURE 10.

Final evidence-informed model of the factors which influence the associations between working from home and health outcomes.