Working From Home, Work-Life Balance, and Depression/Anxiety Among Korean Workers in the COVID-19 Pandemic Period

During the COVID-19 pandemic, working from home is associated with depressive symptoms, and anxiety symptoms in South Korea, and work-family conflict has a mediating role in the associations. These findings were more clearly observed among male, child-bearing, and precarious workers.

Objective

We aimed to investigate the association between working from home (WFH), depression/anxiety, and work-family conflict (WFC) among Korean workers during the COVID-19 pandemic.

Methods

We surveyed a total of 1074 workers online. Depression and anxiety were measured using the Centre for Epidemiologic Studies Depression Scale (CES-D) and Beck Anxiety Inventory (BAI). Mediating effects of WFC on the relationship between WFH and depression/anxiety were examined.

Results

The WFH group had higher depression and anxiety scores than the daily commuting group. As WFC increased, the CES-D and BAI scores also increased. A possible mediating effect of WFC on the relationship between WFH and high CES-D and BAI scores was found.

Conclusion

We observed a significant difference in depression/anxiety between WFH and daily commute workers, which was mediated by WFC, especially for young, child-growing, and precarious workers.

CME Learning Objectives

After completing this enduring educational activity, the learner will be better able to:

• Outline and discuss the association between working from home (WFH), depression/anxiety, and work-family conflict (WFC) among Korean workers during the COVID-19 pandemic

• Outline the difference in depression/anxiety between WFH and daily commute workers

• Discuss the effects of WFC on the relationship between WFH and depression/anxiety

Working from home (WFH) is a working model in which workers do their job from home. Amid the COVID-19 pandemic, WFH increased steeply from 8.2% of US workers in February 2020 to 35.2% in May 2020.¹ In South Korea, WFH also significantly increased from 0.5% of Korean workers in 2019 to 2.5% in 2020 and 5.4% in 2021, although it was lower than that of the United States.² The recent spread of WFH depends on the current workers’ digital mindset as well as on government policies, the development of technology and cloud-based services. WFH is expected to continue after the COVID-19 pandemic since companies have already established systems for WFH, and workers, especially young workers, prefer WFH.³ Workers in the United States expected that WFH would persist after COVID-19 and that 20% of the entire workforce would be supplied by WFH.⁴

Work-family conflict (WFC) can be defined as inter-role conflict, in which the role pressures from the work and family domains are mutually incompatible in some respects.⁵ The detrimental health effects of WFC have been studied, such as self-rated health,⁶ psychosomatic symptoms,⁷ sleep,⁸ health-related behavior,⁹ and health services utilization.¹⁰ Depression and anxiety have also been suggested as consequences of WFC.¹¹

During the COVID-19 pandemic, national containment policies have encouraged employees to work from home for their health and safety. Many companies believe that WFH will become common even after the pandemic and have already set up WFH systems that can decrease operational costs by reducing office space. A sudden shift to WFH could increase workers' difficulties, such as conflicts between work and family.¹² It has been reported that WFH can induce an imbalance between work and family life via cognitive overload because of increased choices and decisions.¹³ There have been several studies on the association between WFH and WFC and the association between WFC and mental health, implying a possible link between WFH and mental health through WFC. However, no studies have addressed whether WFC, a shared variable, is a mediating factor between WFH and mental health. Therefore, this study aims to investigate the association of WFH with depression and anxiety and the mediating role of WFC using cross-sectional survey data of 1074 workers in South Korea.

MATERIALS AND METHODS

Study Population

From August 26, 2021, to September 17, 2021, we conducted a cross-sectional, web-based survey performed by DataSpring Korea, Inc. (https://ko.d8aspring.com/), which provides an online panel survey service. Finally, 1074 participants were recruited, including 533 workers who worked from home and 541 who commuted daily. A remote worker was defined as a worker who had worked from home at least three times a week for the previous 6 months. Survey and informative forms (eg, names of researchers and their institutions, scope and purpose of the study, participation criteria, data privacy commitment form, survey instruments, etc.) were then transferred to an online questionnaire. All responses were anonymous, and no personal identifiable information was requested. Participants were eligible to participate if they¹ were white-collar wage workers aged 20 to 59 years,² had worked from home (4 or more days a week) for the previous 6 months, or³ had commuted daily in the last 6 months. Those with less than 1 year of employment at the current workplace were excluded. The institutional review board of Seoul National University Hospital approved this study (C-2107-253-1242).

Working From Home

Two items were used to measure whether participants engaged in WFH. The first item was “Do you continuously engage in WFH in the last 6 months?” Responses to this item were either “yes” or “no.” If the participants answered “no,” they were classified as daily commuting workers. If they answered “yes,” the following question was asked: ‘If you worked from home, how many days per week did you work from home on average?’ Participants who answered that they had engaged in WFH for more than 4 days or had engaged in WFH and only gone to work when necessary were defined as WFH.

Depression, Anxiety Disorder

The Centre for Epidemiological Studies Depression Scale (CES-D) was used to measure depression. The CES-D was developed by the National Institute of Mental Health to evaluate depression in the general population and to screen for depressive disorders, and it is widely used in epidemiological research worldwide. In Korea, Shin et al¹⁴ modified the items to match the verbal expression of depressive feelings and depressive symptoms of Koreans through a preliminary study.¹⁵ The Korean version of the Centre for Epidemiologic Studies Depression Scale (K-CES-D) consists of 20 items, including depressive affect, positive affect, somatic/retarded activities, and interpersonal factors. The questionnaire consisted of 20 questions, and each question was scored from 0 to 3 points, ranging from a minimum of 0 to a maximum of 60 points, with a higher score indicating a higher level of depression. We used 21 points as a suggested cutoff point for K-CES-D.¹⁶

The Korean version of the Beck Anxiety Inventory (BAI), developed by Beck et al. and translated by Kwon, was used to evaluate anxiety levels in office workers.^17,18 It consists of 21 items, each scored from 0 to 3 points, ranging from a minimum of 0 to a maximum of 63. A higher score indicates a higher level of anxiety. Participants who scored ≥22 were classified as having anxiety. The Korean version of the BAI has been validated.¹⁹

Work-Family Conflict

WFC was evaluated using the work-life balance scale.²⁰ The WFC Scale is an eight-item scale used to measure how work interferes with family in this study. These questions are listed in Supplementary Table S1 (http://links.lww.com/JOM/B218). The items were rated on a seven-point Likert scale ranging from 0 (not at all) to 6 (very much). The final score was obtained through the average response score for each question, with higher scores indicating higher WFC. Cronbach's alpha demonstrated the high reliability of the scale (α = 0.74).

Other Variables

Demographic variables, including age, gender, occupation, education, income, and marital status, were obtained. The number of children was classified into “one or more child” and “no child.” The type of employment was defined using the question “Is your current job (employment) regular or nonregular?” Respondents who answered “regular” were defined as “regular worker,” whereas those who answered “nonregular” were defined as “nonregular worker” workers.

Statistical Analysis

Descriptive statistics were expressed as frequencies and percentages. The χ² test was used to compare the differences between daily commuting and WFH. We calculated the mean and standard deviation (SD) of WFC scores, CES-D, and BAI according to WFH and daily commute work. Subsequently, we performed Student's t test to compare the differences in WFC scores, CES-D, and BAI. After exploring the associations between WFH and mental health (depression and anxiety), a mediation analysis was conducted with WFH as the independent variable, WFC as the mediating variable and mental health (depression and anxiety) as the dependent variable.

To test if the WFC level mediated the effect of WFH on depression and anxiety, we conducted a mediation analysis using the PROCESS macro version 4.0 in SAS,²¹ with multivariate linear regression models at each step (Supplementary Fig. S1, http://links.lww.com/JOM/B218). This approach involved four steps: in the first linear regression model, the correlation between the exposure (WFH) and the outcome (CES-D score, BAI score) was tested, and in the second model, the correlation between the exposure and the mediator (WFC) was tested. The third model tested the outcomes against both exposure and mediators as covariates. Finally, we extracted the total, direct, and indirect effects. In all models, we adjusted for all covariates mentioned above. In the present study, the 95% confidence interval (CI) of the indirect effects was obtained using 5000 bootstrap resamples. For the indirect effects, if zero was not included in the 95% CIs of 5000 times bootstrap tests, a significant indirect effect was concluded. The proportion explained by indirect effect among the total effects was calculated as follows (a × b/c).

We conducted stratified analyses to examine whether the direct effect of WFH and the indirect effect of WFC changed according to age, gender, number of children and type of employment. In addition, we performed an analysis using depression and anxiety as dichotomous variables. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC), and statistical significance was defined as two-sided P ≤ 0.05.

RESULTS

Table 1 presents the demographic characteristics. Of the 1074 participants, 541 commuted daily, and 533 worked from home. Compared with daily commuters, those WFH were more likely to be female, younger, single, and with a higher education level. In addition to these demographic differences, the participants were generally richer and had no children in their WFH space. Compared with daily commuters, those WFH were more likely to be nonregular and white-collar workers.

TABLE 1.

Descriptive Statistics of the Subjects

	Total	Daily Commute	WFH
	N (%)	n (%)	n (%)	P
Total	1074 (100)	541 (50.4)	533 (49.6)
Gender
Male	504 (46.9)	281 (51.9)	223 (41.8)	0.0009
Female	570 (53.1)	260 (48.1)	310 (58.2)
Age, y
20–29	200 (18.6)	107 (19.8)	93 (17.5)	<0.0001
30–39	453 (42.2)	187 (34.6)	266 (49.9)
40–49	281 (26.2)	163 (30.0)	118 (22.1)
50–59	113 (10.5)	68 (12.6)	45 (8.4)
60–	27 (2.5)	16 (3.0)	11 (2.1)
Education
≤ High school	122 (11.3)	77 (14.2)	45 (8.4)	0.0006
College or University	816 (76.0)	411 (76.0)	405 (76.0)
Graduate school	136 (12.7)	53 (9.8)	83 (15.6)
Monthly income (KRW, million)
First quartile	259 (24.1)	130 (24.1)	129 (24.2)	0.5946
Second quartile	278 (25.9)	149 (27.5)	129 (24.2)
Third quartile	253 (23.6)	126 (23.3)	127 (23.8)
Fourth quartile	284 (26.4)	136 (25.1)	148 (27.8)
Marital status
Single	543 (50.6)	263 (48.6)	280 (52.5)	0.1488
Married	492 (45.8)	262 (48.4)	230 (43.2)
Separated, widowed or divorced	39 (3.6)	16 (3.0)	23 (4.3)
No. children
≥1	426 (39.7)	230 (42.5)	196 (36.8)	0.0545
0	648 (60.3)	311 (57.5)	337 (63.2)
Type of employment
Regular	933 (86.9)	492 (90.9)	441 (82.7)	<0.0001
Nonregular	141 (13.1)	49 (9.1)	92 (17.3)
Occupation
White collar	847 (78.9)	371 (68.6)	476 (89.3)	<0.0001
Blue collar	227 (21.1)	170 (31.4)	57 (10.7)

Table 2 shows a significant difference in WFC, depression and anxiety between daily commuters and those WFH (P < 0.05). Generally, the scores of WFC, CES-D, and BAI of WFH participants were greater than those of daily commuters. The result is given in the form of mean (SE).

TABLE 2.

CES-D, BAI, and WFC According to those WFH

	Total	Daily Commute	WFH
	Mean (SD)	Mean (SD)	Mean (SD)	P
Work-life balance
WFC	2.77 (1.05)	2.68 (2.60)	2.86 (2.77)	0.0048
Depression
CES-D	16.81 (12.89)	15.79 (14.75)	17.85 (16.70)	0.0089
Anxiety
BAI	12.58 (12.79)	11.26 (10.30)	13.91 (12.72)	0.0007

Mediation analysis indicated that WFH was positively associated with WFC and BAI (a = 0.2258, P < 0.001 and c’ = 1.7951, P < 0.1) but not with CES-D (Fig. 1). The results did not indicate a direct effect of WFH on CES-D; however, an indirect effect of WFH on CES-D occurred through WFC (b = 1.4856, Boot SE = 0.4463, Boot 95% CI = 0.6082, 2.3544). In addition, the direct effect of WFH on BAI was significant (b = 1.7951, SE = 0.7189, P = 0.0127), and an indirect effect of WFH on BAI occurred through WFC (b = 1.3322, Boot SE = 0.4093, Boot CI = 0.5423, 2.1443).

FIGURE 1.

The mediation model between WFH, WFC, and mental health indices. (A) the mediation model between WFH, WFC, and depressive symptoms measured by the CES-D; (B) the mediation model between WFH, WFC, and anxiety symptoms measured by the BAI. The models was adjusted for age, gender, occupation, education, income, marital status, number of children and type of employment. ***P < 0.01, **P < 0.05, *P < 0.1.

Table 3 shows that the total effect of WFH on CES-D was statistically significant for young (<40 years), male and both regular and nonregular workers after stratification by age, gender, number of children, and type of employment. The indirect effect of WFH on CES-D through WFC was significant regardless of the number of children and type of employment, and when stratified by age and gender, it was significant only in those who were young (<40 years) and male. The mediating effects of WFC proved to be true with bootstrap CIs that did not contain zero. Work–family confilct had a greater mediating effect in the association between WFH and CES-D for young (<40 years), male, with one or more children and regular workers. WFH was associated with a 2.29- to 8.71-point increase in the BAI score (P < 0.05; Table 4). The WFC level mediated the observed effect for the BAI score by WFH.

TABLE 3.

The Mediating Effects of WFC on the Relationship between WFH and Depressive Symptoms (CES-D) Stratified by Age, Gender, Number of Children, and Type of Employment

	Total Effect		Direct Effect		Indirect Effect		Ratio of Indirect Effect to Total Effect
	Effect Size (95% CI)	P	Effect Size (95% CI)	P	Effect Size (95% CI)	P
Mediation analysis stratified by age
<40	2.95 (0.80–5.09)	0.0071	1.49 (−0.34–3.31)	0.1097	1.46 (0.34–2.61)	0.0134	49.55%
≥40	0.98 (−1.47–3.43)	0.4318	−0.39 (−2.46–1.68)	0.7126	1.37 (−0.05–2.86)	0.0568	139.54%
Mediation analysis stratified by gender
Male	3.63 (1.38–5.87)	0.0016	1.57 (−0.27–3.42)	0.0941	2.05 (0.73–3.39)	0.0026	56.61%
Female	1.45 (−0.85–3.74)	0.2168	0.40 (−1.60–2.39)	0.6961	1.05 (−0.09–2.23)	0.0778	72.59%
Mediation analysis stratified by number of children
≥1	2.41 (−0.01–4.82)	0.0507	0.63 (−1.36–2.61)	0.5358	1.78 (0.33–3.31)	0.0139	74.01%
0	2.12 (−0.03–4.27)	0.0536	1.00 (−0.85–2.86)	0.2894	1.12 (0.06–2.22)	0.0488	52.68%
Mediation analysis stratified by type of employment
Regular worker	1.82 (0.15–3.49)	0.0331	0.50 (−0.92–1.91)	0.4897	1.32 (0.39–2.25)	0.0046	72.62%
Nonregular worker	6.29 (0.78–11.80)	0.0257	3.13 (−1.56–7.82)	0.1894	3.16 (0.21–6.69)	0.0488	50.23%

Adjusted for age, gender, occupation, education, income, marital status, number of children, and type of employment.

TABLE 4.

The Mediating Effects of WFC on the Relationship between WFH and Anxiety Symptoms (BAI) Stratified by Age, Gender, Number of Children and Type of Employment

	Total Effect		Direct Effect		Indirect Effect		Ratio of Indirect Effect to Total Effect
	Effect Size (95% CI)	P	Effect Size (95% CI)	P	Effect Size (95% CI)	P
Mediation analysis stratified by age
<40	3.60 (1.49–5.72)	0.0009	2.37 (0.48–4.26)	0.0141	1.23 (0.29–2.23)	0.0141	34.26%
≥40	1.83 (−0.70–4.36)	0.1561	0.49 (−1.69–2.66)	0.6612	1.34 (−0.05–2.90)	0.0573	73.41%
Mediation analysis stratified by gender
Male	4.04 (1.73–6.34)	0.0006	2.22 (0.21–4.23)	0.0305	1.81 (0.64–3.07)	0.0029	44.94%
Female	2.50 (0.24–4.76)	0.0305	1.54 (−0.47–3.55)	0.1324	0.96 (−0.08–2.07)	0.0785	38.32%
Mediation analysis stratified by number of children
≥1	3.76 (1.12–6.40)	0.0053	1.87 (−0.33–4.07)	0.0957	1.89 (0.35–3.57)	0.0141	50.28%
0	2.37 (0.35–4.40)	0.0218	1.51 (−0.33–3.36)	0.1082	0.86 (0.05–1.75)	0.0505	36.22%
Mediation analysis stratified by type of employment
Regular worker	2.29 (0.63–3.95)	0.0068	1.10 (−0.35–2.55)	0.1375	1.19 (0.35–2.06)	0.0047	52.03%
Nonregular worker	8.71 (3.06–14.37)	0.0028	6.03 (0.89–11.16)	0.0218	2.68 (0.17–6.03)	0.0551	30.81%

Adjusted for age, gender, occupation, education, income, marital status, number of children, and type of employment.

After stratification by age, gender, number of children and type of employment, the indirect effect of WFH on BAI through WFC was still significant only in young (<40 years), men, with one or more children and regular workers. In addition, the mediating effects of WFC proved to be true, with bootstrap CIs that did not contain zero. When depression and anxiety were set as dichotomous outcome variables, the results were consistent (Supplementary Fig. S2–3 and Supplementary Table S2–3, http://links.lww.com/JOM/B218).

DISCUSSION

This study demonstrated an association between WFH, depression, and anxiety. WFH is associated with higher WFC and poorer mental health. The results of the present study demonstrate that after controlling for covariates, WFH has an indirect effect on depressive symptoms via WFC but not a direct one. This suggests that WFH was associated with depressive symptoms, but this was because of WFC rather than WFH. In contrast, WFH had direct and indirect effects on anxiety symptoms, which indicates that the association was due to both WFH and WFC. In stratification analyses, the indirect effect of WFH on depression/anxiety via WFC was stronger for young people (<40 years), males, those with children, and nonregular workers. The findings suggest different patterns in the mediating effect of WFC on age, gender, number of children and type of employment. To our knowledge, this is the first study to investigate the impact of WFH and WFC on depressive and anxiety symptoms among South Korean workers.

We found that WFH is related to depressive and anxiety symptoms. These findings are consistent with those of previous studies. A study with a representative sample of 3000 Poles during the COVID-19 restriction period reported that remote work was significantly associated with a deterioration in subjective mental health symptoms.²² A cross-sectional study in Turkey showed protective associations between partial remote work and anxiety levels and between working at the workplace and depression levels but no significant association between remote work and mental health.²³ A Finnish study with 1044 workers during the COVID-19 crisis found that WFH is a significant predictor of COVID-19-related anxiety.²⁴ Previous studies have focused on social isolation, technology use, and psychological stress because of remote work demands to explain the association between WFH and mental health. In addition, job market change and job instability before and after the COVID-19 crisis can also be related to anxiety,²⁴ and the decreased border between work and life due to WFH could also be an attributable factor to detrimental mental health effects.²⁵

We found a significant association between WFH and WFC in Korean workers. These results are consistent with those of previous studies. A recent study examined 318 workers in the Philippines and reported a significantly worsened work-life balance while workers started WFH.²⁶ A study with 299 remote workers reported that increased WFC is associated with decreased work–family integration and argued that WFH induces workers to overwork and disturbs their family role.¹² An Italian study using 211 technical-administrative staff in hospitals who worked from home in April 2020 reported that cognitive demands and technology overload are associated with WFC.²⁷

However, some studies have obtained results different from ours. Ipsen et al. surveyed 5748 European workers who experienced WFH in the COVID-19 pandemic era about the advantages and disadvantages of WFH and concluded that better WLB, better work efficiency, and greater work control were the main advantages.²⁸ Madsen surveyed 98 employees WFH at least 2 days per week and 123 worksite employees and reported that the WFH group showed significantly lower WFC.²⁹ Some studies report that WFH reduced WFC or improved work-life balance,³⁰ whereas others reported opposite results.³¹ These inconsistencies across studies imply that the effects of WFH on WFC could differ by population and situation. In particular, the results of studies conducted during the COVID-19 pandemic can vary because of measures affecting WFC³² and WFH,³³ such as various government regulations like social distancing. Most countries adopted various strategies to prevent the spread of COVID-19, and the strategy imposed by the government included social distancing and lockdowns, which affected not only people's work but also their work-life balance.³⁴ In this regard, the boundary between the workplace and the family area became blurred, especially for workers who did not voluntarily work from home.³⁵ Previous studies have shown that WFH has potential advantages, such as using flexible working days and balancing household and workplace demands.³⁶ However, a greater risk that work spill over into the home has also been revealed.³⁷ Thus, people who work from home may experience conflicts between different roles in situations where work negatively interferes with family duties. For example, staying at home 24 hours during lockdowns can increase the obligations experienced at home (eg, childcare and housework), negatively hindering work demands, leading to WFC. Sandoval-Reyes et al²⁴ studied 1285 remote workers in Latin American countries and reported that remote work demands could be negatively associated with work-life balance and work satisfaction via work stress.

Our study highlighted the mediating role of WFC in the relationship between WFH and mental health (anxiety and depression). Previous studies have reported a relationship among WFC, anxiety and depression. Frone³⁸ analyzed nationally representative data of 2700 workers in the United States and reported that increased WFC was significantly associated with mood and anxiety disorders. Other studies with 220 correction officers, WFC has been reported to be a significant factor for depression.³⁹

Bergs et al⁴⁰ investigated the Maastricht Cohort Study with 3006 participants and reported a significant reciprocal association between depressive symptoms and WFC. A cross-sectional study in Switzerland with 4371 workers reported that WFC was associated with depressive feelings, particularly among female workers.⁴¹ Demands for roles in both work and life domains with limited resources can lead to stress and strain in workers,^42,43 and difficulty in fulfilling domestic obligations and concentrating on work could lead to significant psychological stress. These factors are related to multiple overloaded roles, which could result in psychological work-life conflict and stress.⁴⁴

The results of the stratified analyses imply that the association of WFH with depression and anxiety and the mediating role of WFC could be more apparent among young workers (under 40 years old), males, those who have a child and nonregular workers. Small, confined living spaces can be a stressor for young remote workers. Young workers in South Korea usually live in smaller spaces,⁴⁵ which could be a reason for their vulnerability. Recently, a study in China reported that small housing areas could be associated with depression,⁴⁶ and younger adults may be vulnerable to workplace changes or social isolation stress. One study investigating 5158 Australian adults during the COVID-19 pandemic showed that young age was significantly associated with negative emotions, such as depression, stress and anxiety.⁴⁷ Child-bearing workers could have multiple overloaded roles due to parenting, caregiving or domestic matters, resulting in WFC and psychological stress.⁴⁴ It has also been studied that nonregular workers, who can be classified as precarious workers, are more vulnerable to depression.⁴⁸ The social context of WFH could differ by employment status in the COVID-19 era as WFH for permanent job workers may mean an improved and advantageous working environment; however, WFH for precarious job workers may indicate a more vulnerable employment status because they can be fired more easily than on-site job workers.

Although we expected WFH to have a protective effect on mental health, this study found a negative association between the two. Overall, as the physical boundary between work and life disappears, it becomes ambiguous, and spillover between work and life can follow, leading to WFC. This includes various dimensions; therefore, further research is required. Workers who are married and have a child can be in a stressful condition by longer contact time and conflict during WFH because of house chores and childcare. Workers who live alone must spend a great deal of time in their small living space while performing WFH during the COVID-19 pandemic; thus, their situation can be similar to that of institutionalized people. For vulnerable workers, WFH is completely different from that of stable workers, who experience WFH as a benefit and welfare. However, vulnerable workers may associate WFH with easy layoff and communication only via e-mail or video without in-person relationships. These factors amplify anxiety regarding unstable employment. However, few studies have investigated this issue, and further research should be conducted.

To our knowledge, this is the first study to investigate the relationship between WFH, WFC, depression and anxiety. The association of WFH with depression and anxiety and the mediating role of WFC were the novel findings of this study, which, however, has several limitations. First, the cross-sectional design cannot determine a causal relationship between remote work and mental health. Although we believe that depressive symptoms may not cause a long commuting time, reverse causality should be excluded in a longitudinal study. Second, many individual risk factors for depressive symptoms could not be considered in our study, although our statistical analysis was adjusted for age, region, weekly working hours, income, occupation and shiftwork. Finally, commuting modes, including cars, public transit and walking, were not accessed in the KWCS data. The health effects could vary by commuting mode, although psychological stress can be induced by driving a car and taking a train to work.⁴⁹ Nevertheless, our study has several strengths. First, our results were derived from nationally representative data of the working population of South Korea. Second, we used the WHO-5 index, a validated tool for depressive symptoms that offers good sensitivity and specificity.

CONCLUSION

Compared with on-site work, WFH is significantly associated with depressive and anxiety symptoms among South Korean workers, and WFC could be a mediating factor between these associations. The COVID-19 pandemic has rapidly changed the way of working, but the system and policy were not prepared for WFH. As WFH is gradually spreading, programs and support for WFC and the mental health of WFH workers are needed, especially for young, child-growing, and precarious workers.