The impacts of workplace bullying, emotional exhaustion, and psychological distress on poor job performance of healthcare workers in Thailand during the COVID-19 pandemic Open Access

The rise in COVID-19 cases in Thailand has resulted in a shortage of hospital beds (Haq et al., 2022). The World Health Organization (WHO, 2021) reports that the number of coronavirus cases began to surge in April 2021, with 81,073 new confirmed cases in the final week of July 2021. There have been 578,375 confirmed cases and 4,679 deaths in Thailand since January 2021. In response to increased hospitalizations, 15,960,778 vaccine doses have been administered. Regrettably, the health system is insufficient to address unforeseen events such as the COVID-19 pandemic and future health crises (Tejativaddhana et al., 2022). Therefore, it is crucial to investigate the mental health of Thai healthcare professionals during the COVID-19 period.

Research on psychological well-being and mental health has provided valuable insights into related issues (Anasori et al., 2020; Srivastava and Agarwal, 2020). For example, Yao et al. (2020) studied the impact of workplace bullying (WB) on emotional and cognitive factors among Chinese Research & Development (R&D) employees and emphasized the importance of considering various psychological and cognitive factors instead of solely emotional exhaustion (EE). Similarly, Anasori et al. (2020) investigated the effects of workplace bullying on emotional exhaustion (EE) and psychological distress (PD) among hotel employees in North Cyprus but did not examine its impact on job performance. Although workplace bullying has been found to contribute to emotional exhaustion among hospitality workers in India (Srivastava and Agarwal, 2020), further research is needed to develop a model based on individual-specific variables that can predict its complete effects on employee performance. Haq et al. (2022) explored COVID-19 burnout (CBO) as a moderator in the context of job satisfaction and knowledge hiding among Thai healthcare workers. However, the relationship among these variables remains poorly understood, and there is a need to examine the mental health of Thai healthcare workers, especially during the peak of the COVID-19 pandemic.

It is important to understand how psychosocial factors such as workplace bullying, emotional exhaustion, and psychological distress affect job performance among Thai healthcare workers during the COVID-19 pandemic. This study aims to achieve three objectives: (1) to explore the relationship between WB and PJP, (2) to investigate the mediating roles of EE and PD in this relationship, and (3) to study the moderating effect of CBO on the relationship between WB and EE.

This research adds to existing knowledge about workplace bullying, psychological distress, emotional exhaustion, and COVID-19 burnout among Thai healthcare workers during the pandemic. Previous studies by Srivastava and Agarwal (2020), Yao et al. (2020), and Anasori et al. (2020) have explored related topic. Additionally, it contributes to the ongoing discussion on the moderating role of COVID-19 burnout among hospital staff, as discussed by Haq et al. (2022), Yildirim and Solmaz (2020), and Yildirim et al. (2021). Our findings suggest that emotionally exhausted workers might have had a limited ability to perform well during the pandemic, despite previous literature suggesting the negative impact of emotional exhaustion on job performance (Akhtar, 2021; Çiftci, 2021; Dodanwala and Shrestha, 2021). The authors believe that this may be attributed to the emergency caused by COVID-19, the relevance to the healthcare industry, and the specific job responsibilities of these workers.

Workplace bullying, which is defined as interpersonal conflict among workers, has become a prevalent issue in the knowledge economy (Yildirim and Solmaz, 2020). As organizations consist of employees with diverse physical and emotional traits, clashes and instances of bullying can arise (Dodanwala and Shrestha, 2021). Although WB indirectly affects employee performance, it directly influences employee behavior (Anasori et al., 2020). The stress caused by WB further deteriorates performance in the workplace (Bhatti and Ahmed, 2021). Khalique et al. (2018) argue that WB negatively impacts the productivity of peers and leads to a decrease in employee efficiency, consequently affecting organizational productivity. Moreover, Khurram et al. (2020) highlight that stress and bullying not only harm employee health, but they also create challenges for managers and reduce organizational commitment among employees. WB is found to have an impact on emotional exhaustion and psychological distress (Anasori et al., 2020; Khan, 2021).

EE is characterized as a state of emotional stress resulting from excessive demands and challenges (Lam et al., 2010), which is a significant aspect of the workplace environment. In a supportive work setting, employees typically effectively manage their job responsibilities. However, in a negative workplace environment, employee mood and behavior can be compromised (Karatepe and Olugbade, 2017). Positive or negative emotions play a crucial role in influencing individual reactions to various events in the workplace (Rosander and Blomberg, 2019). They discovered that the presence of WB increases job overload, thereby increasing vulnerability to EE. This heightened state of exhaustion necessitates the use of additional psychological resources, leading to greater levels of exhaustion among employees (Chen et al., 2019). Moreover, workers’ psychosocial issues can have a more significant impact on their health than their work responsibilities (Misiolek-Marín et al., 2020). Consequently, EE has become a significant factor in high-performance occupations, such as nursing (Labrague et al., 2020), particularly during the COVID-19 crisis. Shkoler and Tziner (2017) reported that WB not only damages self-esteem, reduces interactions with peers, and diminishes performance, but also contributes to employees experiencing EE. Employees who frequently face disrespectful encounters are inclined to withdraw and withhold knowledge (Haq et al., 2022), leading to feelings of exhaustion (Malik et al., 2019). Although the mediating role of EE has been examined in various contexts (Anasori et al., 2020; Akhtar, 2021; Çiftci, 2021), its indirect effect of WB on poor job performance through EE has yet to be investigated in the Thai healthcare sector, particularly during the COVID-19 pandemic.

In this study, PD is defined as a collection of unpleasant cognitive and emotional symptoms that affect an individual’s natural mood (Khan, 2021). WB not only leads to PD but also has significant physical (Chan et al., 2019) and psychological (Dodanwala and Shrestha, 2021) consequences. Employees who experience intimidation exhibit worse mental health than their peers (Attell et al., 2017). Therefore, it is essential for decision-makers to recognize the severe impact of stress on employee performance (Berry et al., 2016). Ferguson (2012) found that workplace bullying, characterized by rudeness and unpleasant behavior from subordinates and peers, contributes to PD, leading to reduced employee performance. Previous studies have shown a positive association between WB and PD (Attell et al., 2017; Bhatti and Ahmed, 2021). Moreover, more frequent incidents of bullying have been found to lead to greater distress among nurses (Berry et al., 2016). Furthermore, Chan et al. (2019) concluded that employees who have experienced WB exhibit higher levels of PD, which can have negative implications for work-related outcomes (Khan, 2021). It is worth noting that healthcare staff have also been found to leave their jobs due to elevated levels of PD during the COVID-19 pandemic (Khan, 2021; Haq et al., 2022). Although underlying relationships have been investigated in previous research, no study has examined the impact of PD on poor job performance during the peak wave of the COVID-19 pandemic in the Thai healthcare context.

In hospital settings, establishing a positive and supportive work environment has become a major concern, as the COVID-19 pandemic has disrupted traditional work dynamics. COVID-19-related burnout can lead to decreased job performance and inefficiencies (Yıldırım and Solmaz, 2020). The WHO has recognized burnout as an occupational phenomenon and included it in the International Classification of Diseases-11 (Misiolek-Marín et al., 2020). According to the WHO, burnout is a syndrome experienced by employees who are unable to effectively manage chronic workplace stress. Additionally, Yeun and Han (2016) explained that WB can lead to acute levels of stress in employees, contributing to job burnout. Negative psychosocial factors have been found to have adverse consequences for workers, including depression, stress, depression, psychosomatic disorders, and burnout (Yildirim and Solmaz, 2020; Misiolek-Marín et al., 2020). However, few studies have examined the moderating role of CBO in the context of the COVID-19 pandemic. For instance, Haq et al. (2022) found that CBO significantly buffers the relationship between job satisfaction, knowledge hiding, and turnover intentions. Similarly, Yang et al. (2021) revealed that job burnout plays a significant moderating role in the relationship between safety compliance, safety participation, and safety outcomes among Chinese mining workers. However, few studies to date have investigated the moderating role of CBO in the relationship between WB and EE.

Affective events theory suggests that emotions play a crucial role in employees’ management of workplace situations. This theory explains the link between employee emotions and their response to dynamic workplace events, which influences their work performance. Such theory and the comprehensive work-stress framework (Lu and Kao, 2002) both suggest that workplace incidents can impact employee emotions, which in turn directly affect employee behavior (Weiss and Cropanzano, 1996). In the context of our study on the impact of the COVID-19 pandemic in the Thai healthcare sector, WB can contribute to EE and PD can lead to PJP.

Based on the preceding discussion, this study posits the following hypotheses and presents a conceptual model, as depicted in Figure 1.

The survey method is an effective means for data collection in large populations. The respondents were healthcare workers, including doctors, nurses, administrative staff, and medical assistants. There were two predefined criteria for respondents to participate in the survey: (i) the respondent performed healthcare duties during the COVID-19 pandemic and (ii) the respondent worked in a hospital for not less than one year to optimize the precision of the response concerning the COVID-19 period. The respondents were categorized based on whether they are affiliated with public or private hospitals (e.g., general hospitals, specialty hospitals) or with university or college medical centers. Data were collected using convenience sampling because of COVID-19 restrictions, with healthcare personnel serving as the individual units of analysis. The questionnaire was developed based on previous measurements of related constructs in the English language and then distributed to approximately 300 respondents. Due to unforeseen events, 263 responses were collected; however, 33 were excluded from further analysis due to invalid or incomplete data. The remaining sample size is sufficient for structural equation modeling (Gammoh et al., 2014).

The collected data were compiled by using Microsoft Excel, and a CSV file was produced, which was then passed into Smart-PLS 3.3.3 for estimating partial least squares-structural equational modeling (PLS-SEM). SEM was applied to estimate the proposed research modeling using a two-stage approach (Hair et al., 2019), comprising a measurement model assessment and a structural model assessment. The measurement model assessment estimates the reliability and validity outcomes of constructs, and the structural model assessment assesses relationships between constructs in the model. To determine the PLS-SEM of the structural model, three factors related to the objectives and the characteristics of the data drove the decision to determine the PLS-SEM of the structural model. First, PLS-SEM is recommended when a study seeks to develop or expand upon existing theories and predict the relationships between constructs (Hair et al., 2019). Second, it is particularly useful for estimating latent variables that cannot be directly observed, such as attitudes, psychological characteristics, and behavior (Gammoh et al., 2014). Finally, PLS-SEM provides increased precision and confidence in estimates, even when dealing with limited sample sizes or data sets that do not conform to specific distributional assumptions (Gammoh et al., 2014).

Workplace bullying was assessed using four items from Escartín et al. (2017), and emotional exhaustion was measured with a nine-item scale by Maslach and Jackson (1981). Four items were used for psychological distress from Lynch et al. (2006) and Mewton et al. (2016). Poor job performance was measured with a four-item scale adapted from Noor et al. (2023) and Strebler (2004). A ten-item scale was used to assess the responses related to COVID-19 burnout (Yildirim and Solmaz, 2020). A five-point Likert rating scale ranging from strongly disagree (1) to strongly agree (5) was used. All the items were adapted, and the adaptation process preserved the overall meaning of the instruments ( Appendix).

This study collected primary data from healthcare workers, including doctors (30 percent), nurses (31.30 percent), administrative staff (23.04 percent), and medical assistants (15.65 percent), among which 57.39 percent were female and the other 42.61 percent were male. A total of 29.13 percent were affiliated with general hospitals, 17.39 percent were with specialty hospitals, 27.39 percent were with government hospitals, and 26.09 percent were with university or college medical centers. It is important to note that the respondents were selected to represent a variety of age groups, educational backgrounds, and levels of professional experience. A comprehensive breakdown of these demographics can be found in Table 1.

The authors began with descriptive statistics, examining the mean and standard deviation. These findings are reported in Table 2. Measures of central tendency (i.e., mean) capture the normality of the data, whereas the standard deviation shows the dispersion of the data from its mean value or variations in the data. Respondents were asked to respond on a scale from 1 (strongly disagree) to 5 (strongly agree). The highest mean and standard deviation values were 4.337 and 1.213, respectively.

To measure the indicator reliability, construct reliability, and convergent validity, the outer loadings (OL), Cronbach’s alpha (α), composite reliability (CR) and the average variance extracted (AVE) were measured and are reported in Table 2. The indicator’s reliability was first measured through outer loadings. Table 2 illustrates that all OL values were above the threshold point of 0.70 (Hair et al., 2019) except for two indicators, EE9 and CBO10, with 0.581 and 0.309 loadings, respectively. Thus, EE9 and CBO10 were removed due to low OL scores to ensure the reliability of the model. Moreover, statistically significant (p<0.05 or **) OL values indicated satisfactory indicator reliability.

CR and α were used to assess the construct’s reliability and internal consistency. All coefficients for CR and α ranged between 0.777 and 0.943 and 0.857 and 0.952, respectively, which were above the threshold point of 0.70 (Hair et al., 2019; Haq et al., 2022). Thus, the reliability of the model was maintained. The convergent validity of the model was measured by the AVE. AVE coefficients were between 0.536 and 0.692. With AVE coefficients above 0.50 for all variables, convergent validity was achieved (Hair et al., 2019).

The heterotrait–monotrait ratio (HTMT) was assessed to determine the discriminant validity of the model, as presented in Table 3. All the values for HTMT were within the acceptable range (below 0.90) and met the defined criteria recommended by Hair et al. (2019) and documented by Annan-Prah and Andoh (2023). All variables were discriminant, and discriminant validity of the model was achieved.

Following the measurement model was the structural model assessment, which estimated the statistical significance of beta (β) path coefficients, predictive relevance (Q²), and explanatory power (R²) were measured following the core criteria in structural model assessment. In considering R², Hair et al. (2019) recommended that coefficients of 0.25, 0.50, and 0.75 indicate weak, moderate, and substantial explanatory power, respectively. Table 3 illustrates that the explanatory power of emotional exhaustion (0.591), psychological distress (0.604) and poor job performance (0.605) was moderate. It confirmed that the model explained 59.1 percent, 60.4 percent, and 60.5 percent of the variance in emotional exhaustion, psychological distress, and poor job performance, respectively.

In assessing the predictive relevance (Q²) of the model, Hair et al. (2019) posited that Q² values greater than zero, 0.25, and 0.50 suggest small, medium, and large predictive relevance of the model, respectively. From Table 4, emotional exhaustion, psychological distress, and poor job performance had large predictive relevance. In sum, the model demonstrated large predictive relevance.

The results of all hypotheses are reported in Table 4. Workplace bullying was found to significantly predict poor job performance. Workplace bullying had a positive significant impact on poor job performance $(β=0.430;t=4.386;p=0.001)$⁠, emotional exhaustion $(β=0.424;t=3.965;p=0.001)$⁠, and psychological distress $(β=0.777;t=19.323;p=0.000)$⁠. Thus, hypotheses 1, 2, and 3 were accepted. Furthermore, emotional exhaustion had an insignificant negative effect on poor job performance $(β=−0.053,t=1.505;p=0.118)$⁠. As a result, hypothesis 4 was rejected. In contrast, psychological distress had a positive significant influence on poor job performance $(β=0.456;t=4.649;p=0.001)$⁠. Thus, hypothesis 5 was confirmed and accepted.

Following the direct path analysis, the results of the moderation analysis hypothesis are shown in Table 4. COVID-19 burnout had a statistically significant and positive moderating effect between workplace bullying and emotional exhaustion $(β=0.271;t=6.809;p=0.001)$⁠. Thus, hypothesis 6 was accepted.

The results of the mediation analysis are also reported in Table 4. The indirect effect of workplace bullying → emotional exhaustion → poor job performance was statistically significant $(β=0.039,t=1.302;p=0.002)$⁠. Likewise, the specific indirect effect of workplace bullying → psychological distress → poor job performance was statistically significant $(β=0.355,t=4.418;p=0.001)$⁠. Thus, hypotheses 7 and 8 were accepted and they demonstrated a partial mediation. Both emotional exhaustion and psychological distress had a complementary partial mediating effect between workplace bullying and poor job performance. Comparatively, psychological distress had a stronger mediating role than emotional exhaustion.

Figure 2a presents a graphical representation of the Smart-PLS algorithm function, which illustrates the explanatory power (R²), beta coefficients (β), and factor loadings of the construct indicators. R² values are presented in circles ranging from 0.591 to 0.605, and the latter values are highlighted in yellow, indicating the beta coefficients of each direct path or direct effect. Moreover, values given on the arrows indicate each item’s factor loading. Subsequently, Figure 2b shows the t-statistics for each direct effect and factor loading. Coefficients highlighted in yellow indicate the t-statistics of each path coefficient. Additionally, values given on the arrows indicate t-statistics for each item’s factor loading. This further confirmed that the results presented in Table 4 were aligned with the subsequent graphical representation of the Smart-PLS algorithm (Figure 2a).

The authors utilized simple slope analysis to examine the interaction effect of a third variable, such as COVID-19 burnout, on the relationship between two other variables in our model. Figure 3 presents the results of the simple slope analysis, specifically focusing on the interaction effects of COVID-19 burnout as a moderator. The analysis reveals that the direction of the association between workplace bullying and emotional exhaustion is contingent upon the moderating effect of COVID-19 burnout. Therefore, Figure 3 provides empirical evidence to support the moderating effect of COVID-19 burnout in our model.

Based on the research findings, it is evident that high levels of workplace bullying detrimentally impact healthcare staff, which causes emotional exhaustion, psychological distress, and poor job performance. It aligns with studies by Haq et al. (2022), Khalique et al. (2018), and Curado and Santos (2021), which highlighted the negative consequences of workplace bullying on job performance and turnover. This observation is not surprising, as psychologically distressed employees are unable to perform their duties effectively (Anasori et al., 2020). The presence of workplace bullying, emotional exhaustion, and psychological distress results in a significant decline in overall performance, which has severe implications for the health system.

However, it should be noted that the statistically insignificant impact of emotional exhaustion on poor job performance may suggest that employee commitment and satisfaction (Curado and Santos, 2021; Labrague et al., 2020) play a role in restricting poor performance in hospitals. For healthcare workers, their responsibilities extend beyond mere task management; they have a moral and ethical duty to protect public health and save lives (Nittayasoot et al., 2021). It can be explained as being a “good soldier” or organizational citizenship behaviour (Clercq et al., 2018), they showed positive work behaviour that is not required by formal job description, and engaged in voluntary work efforts regardless of their emotional state. These findings also highlight the fact that healthcare workers in Thailand continue to display high levels of job engagement, even when experiencing emotional exhaustion during the COVID-19 pandemic. This dedication has contributed to Thailand’s success in controlling death rates and new cases, as Thailand was the first country outside China to report a positive case of COVID-19 (Nittayasoot et al., 2021). Additionally, the fear of job loss (Chen and Eyoun, 2021) has incentivized workers to persevere, even when they are emotionally exhausted due to the impact of COVID-19. The significant mediating role of psychological distress is also consistent with previous research (Khan, 2021). COVID-19 burnout among healthcare workers has intensified the impact of workplace bullying on emotional exhaustion, suggesting that a rise in COVID-19-specific burnout is associated with increased emotional exhaustion among hospital staff.

The theoretical implication of this research is twofold. First, scholars are more concerned with a traditional performance construct; studies are still lacking on the poor job performance construct, as performance and poor job performance differ in nature (Noor et al., 2023). Examining the role of the mental health of healthcare workers on poor job performance adds knowledge to the existing literature on affective events theory and the comprehensive work-stress framework. From the COVID-19 perspective, exploring the relationship between negative psychological factors and poor job performance in Thailand using underlying theories, expands potential research avenue and enriches the study of mental health and poor job performance concerning the global health crisis.

The following policy recommendations are suggested. First, exploring the impact of mental issues on poor job performance can guide the healthcare sector to optimize national health policies and improve public health service practices. Policymakers should understand that the higher the healthcare staff’s well-being, the more they tend to improve performance will result in positive supportive actions to public services. For instance, guidelines regarding compensation for adverse events in Article 6(4) of the Finance Ministry 2018 are limited to financial compensations and the Department of Mental Health only offers telephonic counseling for simple illnesses due to a lack of trained staff (Udomratn, 2007). Therefore, the Thai government should prioritize mental health and counseling of healthcare workers by introducing nationwide counseling and training programs to have a sufficiently trained workforce for effective counseling and mental health support of workers with mental health issues. Second, despite a long struggle to implement universal health coverage, the Thai government is still lacking in its requirements set to meet Sustainable Development Goals (SDG) 3.c, a target of 44.5 doctors, nurses, and midwives required per 10,000 population (Nittayasoot et al., 2021). As per WHO recommendation, an optimal workforce density is one trained epidemiologist for a 200,000 population. However, having 183 trained epidemiologists in Thailand equivalent to 0.55 per 200,000 population makes it more challenging and overwhelming to work in the healthcare sector. The problem of insufficient workforce is of utmost importance to address the COVID-19 situation and any future pandemics or health crises. The Ministry of Public Health (MoPH) needs to hire more human resource to align public health policies with SDG 3.c and WHO recommendations for successful implementation through training public health officers and nurses, which will lead to an optimal workforce density to prevent healthcare workers from poor psychological health, resulting in poor performance and undermining the effectiveness of health policies. This study suggests that the MoPH should increase health financing and fund allocation in on-the-job training and development programs to have sufficiently trained epidemiologists, extensive care nurses, and critical care experts to prevent workers from mental health issues due to overwork and enabling them to tackle the current and extraordinary circumstances in the future.

Given the COVID-19 health crisis, it is important to examine whether negative psychological factors have impacted the job performance of healthcare workers in Thailand. The study hypothesizes that workplace bullying is a contributing factor to the emotional exhaustion and psychological distress experienced by healthcare workers during the pandemic. This, in turn, directly or indirectly intensifies poor job performance in the global health crisis context. Furthermore, even in the face of emotional exhaustion, healthcare workers may still perform at work. This finding adds to the literature that emphasizes the detrimental impact of emotional exhaustion on job performance. Importantly, our study highlights that after experiencing extreme burnout specific to the COVID-19 pandemic, workplace bullying becomes an even stronger predictor of emotional exhaustion.

This study has several limitations. First, our findings may not be generalizable to the whole population of healthcare workers in Thailand, or to other sectors and countries. Moreover, during the COVID-19, hospitals in Thailand faced constraints to providing normal healthcare services to the public, which creates uncertainty regarding the structural relationships among the constructs examined in our study in normal situations or under other circumstances. This study suggests that healthcare workers were unable to discontinue their daily job duties during the COVID-19 pandemic, even when experiencing emotional exhaustion. This raises questions about the role of emotional exhaustion, as existing research (Çiftci, 2021; Reb et al., 2017; Wang et al., 2021; Akhtar, 2021) widely supports the notion that emotional exhaustion leads to increased turnover intentions and decreased employee performance. Therefore, further research may investigate whether healthcare workers can effectively manage emotional exhaustion without compromising their work performance. Similarly, it is worthwhile to study COVID-19 burnout as a potential moderator between workplace bullying and psychological distress to gain a deeper understanding of the role of COVID-19 burnout in this relationship.

Funding: The authors would like to extend their sincere appreciation to the National Social Science Foundation (21BGL047), China.

Table A1