Loneliness without distress, chosen isolation (solitude) at the workplace, and mental health and job performance: a cross-sectional study of Japanese employees

Abstract

While loneliness and social isolation in the workplace affect the mental health and job performance of employees, the effects of loneliness without distress and solitude (i.e., chosen isolation) on these outcomes are unclear. The cross-sectional association was examined by using online survey of full-time employees in Japan (n=846). The results showed that the “loneliness with distress” group had significantly higher psychological distress and lower job performance than the other groups. Work engagement was lower both in the “loneliness with distress” and “loneliness without distress” groups, compared to the “non-loneliness” group. The “unchosen isolation” and the “solitude” groups had poorer scores of psychological distress, work engagement, and work performance, compared to the “non-isolation” group. The preliminary findings showed that loneliness without distress and solitude were associated with poor levels of mental health and job performance and should become a target of mental health promotion interventions in the workplace.

Social isolation (the absence of social contact) and loneliness (a feeling of being alone or isolated) have recently received more attention during the COVID-19 pandemic, as loneliness is becoming increasingly common worldwide¹⁾. Loneliness is also common in the workplace. In the UK, 11% of workers often or always experience aspects of loneliness at work²⁾. Social isolation and loneliness in the workplace are an important agenda for occupational health because of their adverse consequences for physical and mental health, as well as job performance³⁾. The negative consequences, including reduced job satisfaction, creativity, and productivity and increased risk of turnover, can be costly⁴⁾.

Loneliness is affected by personal psychological factors⁵⁾. However, work characteristics (such as work hours and workload) and organizational factors (such as organizational climate and organizational support) are associated with loneliness in the workplace^{6, 7)}. However, the issue of heterogeneity in social isolation and loneliness in the workplace has not been explored. First, loneliness, an emotional reaction associated with social isolation, is often associated with psychological distress, such as depression and anxiety⁸⁾, but may not be always so. A proposed theoretical model of loneliness based on plausible psychological and neurobehavioral mechanisms argues that non-distress loneliness may be a transient state leading to loneliness with distress⁹⁾. Is loneliness harmful even when a worker does not feel distress from it? This proposition has not been proven empirically. Second, self-chosen social isolation, called solitude, is usually seen as a self-fulfilling experience for an individual⁸⁾. Is self-chosen social isolation a preferable state? These questions are crucial to identify targets of mental health promotion interventions for social isolation and loneliness in the workplace.

This cross-sectional study thus preliminarily examined the effect of different types of social isolation and loneliness in the workplace, especially loneliness without distress and solitude (i.e., chosen isolation), on mental health and job performance among Japanese employees.

This cross-sectional study used data from the Employee Cohort Study conducted during the COVID-19 pandemic in Japan (E-COCO-J). Full-time employees were recruited from the registered panel of the Japanese online survey company. The baseline data of this cohort was obtained in March 2020 (n=1,448), and the survey was administered repeatedly to those participants who completed the baseline. This study retrieved the data from wave 14 of the survey in September 2023 (n=921; response rate 63.6%). Participants currently working and without missing covariates were included in the analysis. The wave 14 survey was primarily conducted for this analysis. The Research Ethics Committee of the Graduate School of Medicine/Faculty of Medicine, The University of Tokyo No. 10856-(2)(3)(4)(5), approved this study.

Loneliness with/without distress in the workplace was asked a single question: “Loneliness is the subjective feeling of being ‘alone’. You may feel lonely even when you are connected to others. In your workplace or work situation, which of the following applies to you?” The four response options were “I never feel lonely at work” (non-loneliness), “I feel lonely, but do not feel anxious, worried, or stressed about it” (loneliness without distress), “I feel lonely and thus feels anxious, worried, or stressed” (loneliness with distress), and “other”. We excluded participants who responded “other” from the current analysis (n=1).

Social isolation in the workplace and solitude was asked another question: “Isolation refers to having few actual connections with others. In your workplace or work situation, which of the following applies to you?” The three response options were “I am not isolated” (non-isolation), “I am isolated at work, because I choose it but” (solitude), and “I am isolated at work, but it is not what I choose” (unchosen isolation). The items of loneliness and social isolation were only collected at wave 14 survey of E-COCO-J. These items have not been validated in the previous studies.

Psychological distress during the past 30 d was assessed using Kessler-6 (K6) items. Each item was assessed on a scale ranging from 0 (none of the time) to 4 (all the time), with the total score ranging from 0 to 24. The Japanese version of K6 was validated¹⁰⁾, and the Cronbach’s alpha coefficient for the scale in this study was 0.93.

Work engagement during the past 30 d was evaluated using the 9-item Utrecht Work Engagement Scale (UWES-9). All items were assessed on a scale from 0 (never) to 6 (every day). The score was calculated by adding the total scores of all nine items and dividing by 9. Hence, the total score ranged from 0 to 6. The Japanese version of UWES-9 was validated¹¹⁾, and Cronbach’s alpha coefficient for the scale in this study was 0.96.

Job performance over the past four weeks was evaluated using one item of the WHO Health and Work Performance Questionnaire (HPQ). Response options ranged from 0 (worst) to 10 (best). The Japanese version of HPQ was validated¹²⁾.

As another indicator of diminished job performance, counterproductive behavior in the workplace¹³⁾ in the past month was measured using a 10-item original scale derived from a 42-item Japanese scale of organizational dysfunctional behavior¹⁴⁾. Response options were “Yes (1)” or “No (0),” and the total score was calculated by summing the ten scores; therefore, it could range from 0 to 10. The ten items covered a wide range of counterproductive behaviors in the workplace during work hours, such as sending private emails, watching a website for private use, starting work late, leaving work early, breaking the company rules, destructing property, using property for personal gain, ignoring supervisor’s instructions, and the like. Cronbach’s alpha coefficient for the scale in this study was 0.91.

The sociodemographic characteristics included gender, age, educational attainment, marital status, occupation, industry, and company size.

The groups classified based on questions of loneliness and social isolation in the workplace were compared on psychological distress, work engagement, job performance, and counterproductive behaviors. The analysis of covariance (ANCOVA) was used to test statistically significant differences, adjusted for gender and age (in years). Statistical significance was set as a two-sided p<0.05. SPSS 28.0. Japanese version (IBM) was used.

This study included 846 employees. Table 1 presents the participants’ characteristics. Most participants were over 50 yr old (45%), non-manual workers (58%), and employed by a company with over 1,000 employees (33%).

Table 1. Participants’ characteristics (N=846).

		n	%	Mean	SD
Gender
	Men	459	54.3
	Women	387	45.7
Age				46.6	10.1
	20–29 yr	27	3.2
	30–39 yr	201	23.8
	40–39 yr	238	28.1
	Over 50 yr	380	44.9
Educational attainmenta
	Highschool	198	23.4
	College	188	22.2
	University	374	44.2
	Graduate school or above	46	5.4
	Other, missing	40	4.7
Marital status
	Single/divorced/widows	377	44.6
	Married	469	55.4
Occupation
	Manager	96	11.3
	Non-manual workers	487	57.6
	Manual workers/others	263	31.1
Industryb
	Public sector	78	9.2
	Manufacturing	204	24.1
	IT service	81	9.6
	Retail and wholesale business	92	10.9
	Finance, insurance, real estate	59	7
	Professional and technical services	59	7
	Restaurant, hotel business	13	1.5
	Life-related services and entertainment	53	6.3
	Education and learning support	43	5.1
	Medical and welfare	105	12.4
	Transportation/construction	53	6.3
	Primary industry	2	0.2
	Others	4	0.5
Company sizeb
	Over 1,000 employees	281	33.2
	300–999	132	15.6
	50–299	230	27.2
	<50	200	23.6
	Unknown	3	0.4

^aThe data was collected in May 2020 or August 2020. ^bThe data was obtained in February 2022. We used the data from March 2020 if the participants did not respond in February 2022.

SD: standard deviation.

The three groups classified on the basis of workplace loneliness differed significantly in average scores of psychological distress, work engagement, and job performance (Table 2). In general, the loneliness with distress group and the loneliness without distress group scored significantly lower on these outcomes compared to the non-loneliness group. Work engagement was similar in the loneliness with and loneliness without distress groups and significantly lower compared to the non-loneliness group.

Table 2. The difference in mean scores of outcome variables (N=846).

	Loneliness and distress about it
	Non-loneliness (n=498)		Loneliness without distress (n=116)		Loneliness with distress (n=152)		Group difference
							Crude		Adjusted by age and sex
	Mean	SD	Mean	SD	Mean	SD	F value	p-value	F value	p-value
Psychological distress	3.76	4.86	7.61	5.91	9.22	6.38	79.68	<0.001*	75.43	<0.001*
Work engagement	2.40	1.30	1.65	1.23	1.68	1.20	31.99	<0.001*	34.13	<0.001*
Job performance	6.64	1.77	5.81	1.75	5.14	2.23	45.41	<0.001*	41.35	<0.001*
Counterproductive behavior	3.02	5.09	3.20	4.78	4.20	5.39	3.49	0.031*	2.84	0.059
	Social isolation at work
	Non-isolation (n=557)		Solitude (n=130)		Unchosen isolation (n=79)		Group difference
							Crude		Adjusted by age and sex
	Mean	SD	Mean	SD	Mean	SD	F value	p-value	F value	p-value
Psychological distress	3.93	4.96	9.50	6.00	9.31	6.32	90.64	<0.001*	91.53	<0.001*
Work engagement	2.38	1.29	1.59	1.15	1.40	1.16	40.24	<0.001*	40.54	<0.001*
Job performance	6.58	1.82	5.43	1.94	4.98	2.11	43.62	<0.001*	43.14	<0.001*
Counterproductive behavior	2.96	4.93	4.61	5.58	3.38	5.30	6.19	0.002*	5.49	0.004*

*p<0.05.

SD: standard deviation.

The three groups classified on the basis of workplace isolation also differed significantly in the average scores of psychological distress, work engagement, job performance, and counterproductive behaviors (Table 2). Unchosen isolation and the solitude groups scored similarly on psychological distress and significantly higher compared to the non-isolation group. Unchosen isolation and the solitude groups scored similarly on work engagement and significantly lower compared to the non-isolation group. The unchosen isolation group, followed by the solitude group, scored significantly lower on job performance compared to the non-loneliness group. Counterproductive behaviors were significantly higher in the solitude group compared to the unwanted isolation and non-loneliness groups that scored similarly.

The results support that the non-distress loneliness may be an intermediate state in the transition from the non-loneliness group to the loneliness distress group, as suggested by a previous theoretical model of loneliness⁹⁾. The loneliness with distress group had the highest average scores of psychological distress, which is consistent with previous studies³⁾. The results indicated that workers with non-distress loneliness may have better levels of mental health and job performance compared to those who experience loneliness with distress. However, the non-distress loneliness group may still exhibit worse mental health outcomes and job performance than those without loneliness. Thus, non-distress loneliness group could also be a target of interventions to prevent loneliness in the workplace. The nature of non-distress loneliness should be investigated in a future longitudinal study.

Contrary to a previous notion that solitude is a self-fulfilling experience⁸⁾, the present study found that psychological distress and work engagement were poorer to the same extent in unchosen isolation and solitude groups. Job performance of the solitude group was better than that of the unchosen isolation group but still poorer compared to the non-isolation group. Counterproductive behaviors were higher in the solitude group compared to the unchosen isolation and non-isolation groups, which received similar scores. This unexpected, conflicting result may be because the present study investigated solitude in the workplace rather than solitude in the general life domain, such as the family and the community. Workers are usually required to cooperate with colleagues to get the job done in the workplace. A personal effort to keep solitude may conflict with norms, expectations, and rules in the workplace, which might make workers psychologically distressed and less motivated to work. Such solitude may discourage workers from strictly obeying the rules of the company/organization. Alternatively, a desire to disobey the rules of the company/organization may be an initial motivation to obtain solitude. The solitude group could also be a target of an intervention or support to improve mental health and job performance. However, further study is needed to understand the psychological mechanisms underlying solitude in the workplace and develop effective interventions.

There were several limitations in this study. The cross-sectional nature of the study might have biased the results due to reverse causality. For instance, participants who had poor mental health status may have over-reported their loneliness or isolation. A further longitudinal study is needed. Personality traits or personal work values may have confounded the observed associations. In particular, participants who did not value work may prefer solitude in the workplace and have low work engagement. The questions on loneliness and isolation were not validated. Relevant scales should be developed and fully tested for the reliability and validity. In addition, participants may not have understood the questions on loneliness and isolation. It is theoretically inappropriate to define “chosen” loneliness; however, some may misinterpret chosen isolation as chosen loneliness. Considering these questions, the effects of loneliness and isolation on outcomes should be interpreted carefully. The generalizability of the study’s findings may be limited to full-time Japanese employees, mainly white-collar workers, because the observed associations may vary depending on employment status, work values, and workplace culture of a country. For instance, because Japan has a strong group-oriented collectivism culture, the observed association of loneliness and social isolation with health outcomes may be stronger than in countries with individualism cultures. A cross-cultural comparison study would be desirable. The cohort was originally composed of full-time employees. However, we did not ask their employment status at the wave 14. There may be a small number of the self-employed in the sample, who could have a smaller social network at work but better health and productivity outcomes, leading to the possibility of underestimating the association of the chosen isolation with outcomes.

Disclosure

Conflict of interest

NS received personal fees from Medilio Co., Ltd., outside the submitted work. NK obtained personal fees from the Occupational Health Foundation, SB AtWork Corp, RIKEN, Japan Aerospace Exploration Agency (JAXA), Japan Dental Association, Sekisui Chemicals, Junpukai Health Care Center, Osaka Chamber of Commerce and Industry outside the submitted work. KI and NK are employed at the Department of Digital Mental Health, an endowment department supported with an unrestricted grant from 15 enterprises (https://dmh.m.u-tokyo.ac.jp/c) outside the submitted work. RK received personal fees from SATORI Electric CO., LTD, NXP Semiconductors, RIKEN, Toyotsu Chemiplas, and Mitsubishi Materials Corporation outside the submitted work. All other authors declare no competing interests.

Funding/Support

The Japan Science and Technology Agency (JST) Grant Number JPMJRS22K1, Japan, supported this research.

Role of the funder/Sponsor

The sponsors played no role in the design and management of the study; data collection, analysis, and interpretation; manuscript preparation, review, and approval; and the decision to submit the manuscript for publication.

Ethics approval and consent to participate

The Research Ethics Committee of the Graduate School of Medicine/Faculty of Medicine, The University of Tokyo, No. 10856-(2)(3)(4)(5), approved this study. All methods were carried out in accordance with the Declaration of Helsinki.

Informed consent

Online informed consent was obtained from all participants with full disclosure and explanation of the purpose and procedures of this study. The panelists had the option to not respond to any part of the questionnaire and to discontinue participation in the survey at any point.

Registry and registration number of the study/trial

N/A.

Animal studies

N/A.

Data availability

The data supporting this study’s findings are available from the corresponding author, NK, upon reasonable request.

Author contributions statement

NK was in charge of this study, supervising the process and providing his expert opinion. NS and NK organized the study design and analyzed the data. Collaborators RK, KT, and KI ensured that questions related to the accuracy or integrity of any part of the work were appropriately investigated and resolved. All authors participated in conducting the survey. NS and NK wrote the first draft of the manuscript, and all other authors revised it critically. All authors approved the final version of the manuscript.