Workplace interpersonal relationships and turnover intentions among care workers working in older adult care facilities in Japan: a mediation analysis comparing direct and mediated effects via psychological distress

The older population is growing rapidly globally, and this trend is expected to continue for some time [1]. As a result, the incidence of age-related diseases such as diabetes [2], neurodegenerative diseases [3], cardiovascular diseases [4], kidney disorders [5], and other age-related diseases are increasing. Age-related diseases increase the risk of older people needing long-term care. In Japan, 2.8% of individuals aged 65–69 years need long-term care insurance, compared with 5.8% of those aged 70–74, 12.1% of those aged 75–79, 25.8% of those aged 80–84, 48.5% of those aged 85–89, and 76.8% of those aged ≥ 90. This trend increases with age [6, 7]. This suggests that with the growing older population worldwide, the number of older people requiring care increases. Therefore, the role of care workers in caring for older people will also increase in the future.

The shortage of care workers is a global problem [8, 9], and the turnover rate of care workers is also high [10]. To address the shortage of care workers, it is important not only to increase the number of new care workers but also to stem the turnover of care workers in the workforce. Actual turnover is difficult to control because of factors unrelated to turnover intentions, such as organizational decisions, health problems, and work contract termination [11]. However, turnover intention refers to the attitude or perception of workers that can be changed through organizational intervention [12]. Because turnover intentions are closely related to turnover [13], organizational attempts to lower the turnover intentions of care workers should effectively compensate for the shortage of care workers.

Previous studies revealed that factors associated with care workers’ intentions to leave the workforce include support from supervisors and colleagues [14, 15]; interpersonal conflicts with supervisors, colleagues, and subordinates [16, 17]; perceptions of teamwork in the workplace [18]; affective organizational commitment [19]; employment status and other treatment [20]; job satisfaction [14]; and work burden [21]. In particular, many studies have reported on the relationship between workplace relationships and turnover intentions, and a review of studies of Japanese care workers revealed that workplace interpersonal problems were consistently associated with turnover intentions [22]. Conversely, turnover intentions among care workers are associated with high levels of job stress, including psychological distress [17, 23]. Care workers involved in older adult care are susceptible to psychological distress and burnout [24, 25]. In addition, peer and supervisor support is a predictor of burnout among residential aged care workers [26], and workplace interpersonal problems are also associated with psychological distress among care workers [27].

From the above, it can be inferred that different workplace interpersonal problems faced by care workers working in older adult care facilities have different effects on psychological distress and turnover intentions. If this point is clarified, valuable implications can be identified between workplace interpersonal relationships that should be worked on directly to improve the turnover intentions of care workers and those that should be improved to maintain mental health. Deterioration of care workers’ mental health not only negatively affects the care workers themselves but also decreases the quality of care they provide [28]. In addition, high turnover among care workers not only increases the burden on care workers working in the field but also decreases the quality of care [29]. Therefore, the relationship between workplace interpersonal relationships, psychological distress, and turnover intentions must be examined to reduce the burden on care workers and maintain the quality of care. However, this point has not been adequately examined to date.

This study aimed to investigate the relationship between workplace interpersonal relationships, psychological distress, and turnover intentions among care workers working in older adult care facilities. We hypothesize that some workplace interpersonal problems among Japanese care workers directly influence turnover intentions, whereas others influence turnover intentions through the mediation of psychological distress (Fig. 1). This finding may be useful for considering efficient organizational approaches to prevent turnover intentions among care workers.

Table 1 summarizes the basic attributes of the study participants. The mean age was 44.7 ± 12.1 years. The most common final education was high school (43.7%), and 77.6% of all respondents were certified care workers. The majority of the care facilities where they worked were residential services (68.1%), 91.3% were employed full-time, and 70.1% were employed with no managerial position. The means and standard deviations, Pearson’s correlation coefficients, and ω coefficients for the basic attributes (for scale/subscale measures only) and all scale scores are summarized in Table 2. All scales used in this study showed good reliability [36].

The results of the mediation analysis using total scores for the WIPS are presented in Table 3. In Model 1, workplace interpersonal problems had a significant positive overall effect on turnover intentions (β = 0.520 [95% CI = 0.392,0.624]). Of these, the mediating effect of workplace interpersonal problems on turnover intentions through psychological distress was significantly positive (β = 0.227 [95% CI = 0.129,0.322]). Workplace interpersonal problems also had a significant positive direct effect on turnover intentions (β = 0.293 [95% CI = 0.141,0.446]). Thus, psychological distress partially mediated workplace interpersonal problems. This was also replicated in Model 2 findings, which was adjusted for basic attributes (Total effect: β = 0.528 [95% CI = 0.415,0.634]; Direct effect: β = 0.336 [95% CI = 0.194,0.469]; Mediation effect: β = 0.192 [95% CI = 0.104,0.281]). Notably, the direct effect was more influential than the mediating effect in Model 2.

The results of the mediation analysis for the WIPS subscale are shown in Table 4. In Model 1, all subscales showed significant positive overall effects on turnover intentions. Of these, the direct effect of a sense of unfair workload was stronger than the mediating effect of psychological distress (Direct effect: β = 0.318 [95% CI = 0.166, 0.455]; Mediation effect: β = 0.205 [95% CI = 0.125, 0.298]). Psychological distress partially mediated workplace relationship problems related to insufficient communication (Direct effect: β = 0.284 [95% CI = 0.139, 0.425]; Mediation effect: β = 0.229 [95% CI = 0.138, 0.329]), differences attitudes to care work (Direct effect: β = 0.235 [95% CI = 0.100, 0.364]; Mediation effect: β = 0.232 [95% CI = 0.153, 0.325]), difficulty in guidance for subordinates/new staff (Direct effect: β = 0.179 [95% CI = 0.050, 0.307]; Mediation effect: β = 0.227 [95% CI = 0.147, 0.331]), and labeling (Direct effect: β = 0.184 [95% CI = 0.057, 0.335]; Mediation effect: β = 0.237 [95% CI = 0.160, 0.330]). In the case of bullying, psychological distress completely mediated the relationship as the direct effect was not significant, whereas the mediating effect was significant (Direct effect: β = 0.121 [95% CI = − 0.017, 0.241]; Mediation effect: β = 0.232 [95% CI = 0.159, 0.318]).

The only change in the results of Model 2 compared to those of Model 1 was that the direct effect was stronger than the mediating effect for insufficient communication (Direct effect: β = 0.319 [95% CI = 0.192, 0.459]; Mediation effect: β = 0.194 [95% CI = 0.105, 0.285]). Additionally, bullying was partially mediated by psychological distress and workplace relationship problems (Direct effect: β = 0.207 [95% CI = 0.091, 0.322]; Mediation effect: β = 0.204 [95% CI = 0.135, 0.290]). There were no clear changes compared with Model 1 in the sense of unfair workload (Direct effect: β = 0.314 [95% CI = 0.187, 0.443]; Mediation effect: β = 0.190 [95% CI = 0.113, 0.268]), different attitudes toward care work (Direct effect: β = 0.278 [95% CI = 0.140, 0.401]; Mediation effect: β = 0.204 [95% CI = 0.128, 0.293]), difficulty with guidance for subordinates/new staff (Direct effect: β = 0.207 [95% CI = 0.072, 0.329]; Mediation effect: β = 0.219 [95% CI = 0.141, 0.313]), and labeling (Direct effect: β = 0.198 [95% CI = 0.073, 0.330]; Mediation effect: β = 0.211 [95% CI = 0.143, 0.287]).

To the best of our knowledge, this was the first cross-sectional study that examined the relationship among workplace interpersonal problems, psychological distress, and turnover intentions of care workers working in older adult care facilities. The results showed that workplace interpersonal problems had a significant positive overall effect on turnover intentions. This finding is consistent with previous reports that workplace relationships are associated with turnover intentions [17, 37]. In particular, a study of healthcare workers showed a negative association between workplace social capital and turnover intentions, even after controlling for work-related factors [38]. Workplace social capital includes a vertical dimension (relationship between employees and their employer or supervisor) and a horizontal dimension (social contact, cooperation, and trust among coworkers) [39] and strongly reflects interpersonal relationships in the workplace. Thus, workplace interpersonal problems are an important predictor of care workers’ intentions to leave their jobs. Conversely, some workplace relationship problems had a strong direct impact on turnover intention, whereas others influenced turnover intention primarily through the mediating effect of psychological distress. These findings suggest that different workplace interpersonal problems can be targeted to reduce care workers’ turnover intentions.

First, a sense of unfair workload and insufficient communication had significant positive direct effects on turnover intentions in Models 1 and 2. First, a sense of unfair workload and insufficient communication had large positive direct effects on turnover intentions in Models 1 and 2. The sense of unfair workload suggests dissatisfaction because of differences in workload among care workers, which means that supervisors cannot manage the workload properly and care workers have difficulty cooperating [17]. Satisfaction with supervisors is negatively related to the turnover intentions of care workers [40]. In addition, better leadership increases the affective organizational commitment of care workers [41], and higher affective organizational commitment lowers the turnover intentions of elderly care workers [19]. Thus, the sense of unfair workload held by individual care workers, which is related to their supervisor’s management, appears to have a direct influence on turnover intentions. Conversely, differential attitudes to care, such as conflicting goals, hinder communication [42, 43]. The different attitudes to care work, which were found to have equal direct and mediating effects on turnover intentions in Models 1 and 2, would result in insufficient communication. A study of nurses showed that workplace communication is a protective factor for job retention intentions [44]. Insufficient communication lowers the perception of teamwork in the workplace; however, older adult care workers were found to have higher perceptions of workplace teamwork lower turnover intentions [18]. In summary, insufficient communication can be a direct factor that increases turnover intentions among care workers. These results suggest that improving the sense of unfair workload, insufficient communication, and different attitudes to care work among workplace interpersonal problems can reduce turnover intentions of care workers and may be an effective means of securing the caregiving workforce.

Second, difficulty guiding subordinates/new staff, labeling, and bullying had large mediating effects on turnover intentions through psychological distress in Model 1. In Model 2, the direct and mediating effects of bullying on turnover intentions were similar; however, difficulty in guidance for subordinates/new staff and labeling exerted a significant mediating effect on turnover intentions through psychological distress. Because of the high turnover rates of care workers and the severe labor shortage in Japan [22, 45], new staff are expected to join older adult care facilities each year. In Japan, older adult care facilities allow people to enter the caregiving profession without having specialized knowledge or skills [22]. As a result, senior care workers already working there must provide fairly extensive guidance to new staff and subordinates. In addition to being busy with their daily duties, partly due to manpower shortage, care workers may feel quite overwhelmed in guiding new staff and subordinates. Labeling is a cognitive distortion in which people in the workplace are labeled negatively [17]. Cognitive distortions affect social adjustment through psychological distress such as depressive symptoms [46]. Based on these findings, negative views of workplace staff through labeling may mediate psychological distress, making workplace adjustment more difficult and increasing turnover intentions. In the model adjusted for basic attributes, bullying had equal direct and mediating effects on turnover intentions. This suggests that bullying can be a direct factor in increasing turnover intentions or a mediating factor in psychological distress leading to turnover intentions. To support this finding, a review of nursing staff reported that bullying causes psychological distress and increases the likelihood of isolation from colleagues [47]. Thus, improving workplace interpersonal problems such as difficulty guiding subordinates/new staff, labeling, and bullying is effective in preventing psychological distress among nursing staff and subsequently the intentions to leave the nursing profession. This will also help reduce turnover intentions of care workers in the long run.

This study had some limitations. First, owing to the cross-sectional nature of the study, we could not capture the temporal sequence of the process. It is difficult to confirm causality between workplace interpersonal problems, psychological distress, and turnover intentions using a cross-sectional approach. Although cross-sectional studies using mediation analysis have been conducted in the past [48], researchers should use caution when attempting to use cross-sectional data in place of longitudinal data for mediation analyses [49]. Second, the valid response rate was relatively low (31.3%), and the results of this study may have been influenced by selection and social desirability bias. In particular, those having interpersonal problems and psychological distress are less likely to be willing to answer the survey questions, which causes bias in the estimation results. Furthermore, the small sample size may lead to wide CIs, which should be interpreted with caution. To contribute to policymaking and workplace interventions, future studies need to be conducted with a larger sample size, using a longitudinal survey of care workers in elderly care facilities. Third, a longitudinal study in Northern Europe showed that job dissatisfaction resulting from adverse working conditions is associated with job switching [50]. Patients with dementia can unintentionally behave in ways harmful to care providers [51]. These findings suggest that harms and hazards affect caregivers’ turnover intensions. In this study, we did not collect data on adverse working conditions such as harms and hazards. In future research, it will be necessary to examine the impact of workplace interpersonal relationships on psychological distress and turnover intentions after adjusting for these factors. Fourth, this study focused on care workers’ turnover intensions. However, it is not the case that care workers who intend to leave their jobs actually do so. For this reason, it must be emphasized that the results of this study do not allow us to comment on the impact of workplace interpersonal relationships and psychological distress on care workers’ actual turnover.

This study revealed that some workplace interpersonal problems have a stronger direct effect on turnover intentions, whereas others have a stronger mediating effect on turnover intentions through psychological distress. Workplace interpersonal problems are predictors of turnover intentions among care workers working in elderly care facilities, and different workplace interpersonal problems may be targeted for intervention to reduce turnover intentions. More studies are needed to determine whether improving workplace interpersonal relationships will prevent care workers from leaving their jobs.