Although some progress has been made in recent years, including the improved quality of training programs (1), the psychiatric workforce is still in severe shortage (2) and unevenly distributed in China. Maintaining workforce sustainability and reducing attrition are vital ways to slow down the shortage of mental health workforce (3). To reduce turnover rate for psychiatrists, it is important to identify the upstream drivers of turnover intention of psychiatrists, and then to inform strategies in recruitment and retention.

In recent decades, workplace violence (WPV) against healthcare workers has drawn attention and become a serious public health issue globally (4, 5). Workplace violence can be defined as “incidents where staff are abused, threatened or assaulted in circumstances related to their work, involving an explicit or implicit challenge to their safety, wellbeing or health” (6). According to the World Health Organization (WHO), health workers are subject to high risk of violence, with 8%-38% of health workers suffering from physical violence at some point in their careers, and even more are threatened by or exposed to verbal aggression (7). It has been found that healthcare workers in the psychiatric settings are approximately three times higher of encountering violence (8, 9) when compared to other medical specialties. Patients' violent behavior, especially physical aggression and verbal violence in psychiatric settings, was of great concern for researchers (10–12) and has become increasingly common in China (13, 14).

WPV can physically and psychologically harm healthcare workers and affect the workforce stability (15). A growing body of research literature suggests that WPV is a key risk factor associated with turnover intention. For instance, a study from an emergency department in Korea showed that WPV increased nurses' intention to leave the hospital (16). A study among physicians in China found that WPV was significantly associated with turnover intention (17). Another survey with physicians in Finland also showed that WPV led to increased turnover intention (18).

The stressor–strain–outcome (SSO) model specified a process between stressors and the outcome with a mediating role played by strain (19, 20). In the SSO model, stressors are environmental stimulus that individuals experience, strains are individuals' personal emotions, and outcomes are behavioral responses to stressors (21). According to the SSO model, strain plays an essential role in the associations between workplace violence and turnover intention. Other related research reaches the consensus that stress and exhaustion are viewed as transactional processes between individual environment stimulus and response behaviors (22, 23). A meta-analytic review proposed that, the negative effects of WPV on affects and attitudes would have indirect consequences on behavioral outcomes (24). Many researchers further examined mental health as a mediator in the association of WPV with turnover intention. Recently, burnout and stress were considered as key mediators and have been examined (25–27). Nonetheless, few of the existing studies systematically explored the relationship between WPV and turnover intention among psychiatrists, and none have examined the mechanism embedded in the stressor-to-outcome process.

With a widely recognized shortage of mental health care workers (28), preventing brain drain has become one of workable solutions to psychiatrists shortage both in short-and long-term. Understanding the impacts of WPV on turnover intention is crucial for preventing more psychiatrists from quitting their job. This is of even more significance in China as a high rate of WPV in healthcare setting has been reported (14). Our study, based on the SSO model, would reveal the transactional process among WPV, mental health and turnover intention in order to provide insight for prevention interventions (22). Based on the SSO model, our research examined the transactional process from WPV to turnover intention by focusing on the following: (1) identifying upstream drivers of turnover intention among psychiatrists; (2) examining the association between verbal and physical violence and turnover intention among psychiatrists; (3) exploring the mediating roles of burnout and stress in the process. The study aimed to understand the mechanism behind turnover intention and shed light on policymaking, in order to maintain human resource sustainability in psychiatric settings.

With a large national sample of psychiatrists from 41 tertiary psychiatric hospitals in China, we found that psychiatrists experienced higher rates of verbal violence (78.0%) and physical violence (30.7%) when compared to the overall rates of verbal violence (48.52%) and physical violence (5.84%) in 144 Chinese public hospitals (40). Some possible reasons may explain the high rates. In psychiatric settings, patients with unstable mental status may be stressed and are more likely to express anger toward those who care for them. In addition, due to the increasing demand of mental health services in China, psychiatrists are facing an inflow of patients suffering from mental illness, which would increase their workload. Hospitalizations in psychiatric settings maybe for days, weeks, or longer, and this ongoing care relationship may also increase the incidence of WPV (41–43). The prevalence of WPV vary across socio-economic groups. This finding is consistent with our other study (44), where we found that WPV undermines workforce stability and got inspired to further explore the mechanisms between WPV and turnover intention. In this study, we examined the associations between workplace violence and turnover intention, and the mediating effects of mental health on the associations, contributing to the existing literature on WPV, mental health and turnover intention among physicians. Our findings suggest the existence of a psychologically transactional process (as strains) between WPV (as individual environment stimulus factors) and turnover intention (as response behaviors) among psychiatrists.

We found a high rate (31.9%) of psychiatrists reporting they would leave their job. The percentage was also much higher than the turnover intention of nurses (20.0%) and pharmacists (17.8%) in psychiatric settings (1, 45). Consistent with previous studies (46–48) in several countries, psychiatrists with higher salary and satisfaction with their health conditions are more likely to stay in their current job; male doctors with longer work history and working hours are more likely to have turnover intention. These findings may help policymakers and hospital administrator to take more targeted measures (e.g., health promotion, lower hours of work or higher level of salary) to retain employees and boost their wellbeing. In addition, consistent with previous studies, healthcare workers who reported burnout and stress were more likely to have turnover intentions (49–51). This finding suggests that reducing burnout and stress among psychiatrists may be effective in preventing their turnover intention.

We found that psychiatrists with greater exposure to WPV were more likely to have turnover intention. This is consistent with previous studies in which workplace violence was found to be associated with higher turnover intention among physicians (52, 53). As the experience of verbal violence and physical violence were positively associated with leave intention, it was not surprising that such “negative experience” could contribute to the increasing turnover intention. Concordant with previous studies where physical and verbal violence are found to have significant impact on burnout among Chinese physicians (40), our findings suggest that WPV, including both verbal violence and physical violence, is associated with burnout and stress. A study in Israel also suggests that verbal and physical violence were significantly associated with stress among mental health nurses (10).

Moreover, the mediation analysis in this study showed that the relationship between WPV and turnover intention was mediated by mental health. This result supports the mediating effect of mental health (including occupational stress and burnout) reported in association between patient violence and turnover intention among nurses (25). Our findings were also partly in line with the observation during COVID-19 in China that the effect of workplace violence on turnover intention was partially mediated by mental health (i.e., stress) (27). Although exposure to WPV might lead to mental health problems, our analysis indicates that mental health is a meaningful knob to reduce the negative impact of WPV on turnover intention in this population.

Strategies are needed to reduce and prevent the incidence of WPV (54, 55). Adequate training, education and support should be provided for healthcare workers (56). Improving communication skills of healthcare workers would also be helpful in reducing patient-physician misunderstandings (57). Legislation against WPV in healthcare facilities can be established in order to protect vulnerable healthcare workers (9). Governments could declare a crackdown on assault on doctors and on violence in hospitals (14). Our findings additionally suggest that hospital administrators should take measures to promote the mental health and wellbeing of psychiatrists after the exposure to WPV. This indicates that, besides providing sufficient support, helping psychiatrists deal with mental health in the face of WPV might be equally important to reduce their turnover intention (58). Just as researchers advocated, hospital management should consider the effects of workplace violence on workers' mental health and their subsequent intention to leave (59). The de-escalation of incidents would be an effective way to decrease the likelihood of a violent outcome (60). For example, the hospital management policy should conduct stress reduction intervention program to mitigate mental health burden (10). Hospitals can also provide specialized and confidential trauma-informed mental health services (61) and provide timely psychological interventions (such as establishing professional pressure relief workshops), and arrange psychologists for physicians who have experienced workplace violence (58).

The study has a few limitations. First, the data was collected in a self-reported manner, which might have cognitive bias. Second, the data sets were cross-sectional with no causal relationship can be inferred. Third, psychological violence and sexual harassment and other implicit violence were not included in our survey. While we discussed the mediation effects of burnout and stress, more potential mediators (e.g., job performance, professional identity, and job security) between WPV and turnover intention could be further explored. Fourth, as STATA does not provide GSEM model fit information, we assessed RMSEA and SRMR with a large sample size under the linear assumption instead. Finally, the study focused primarily on violence from patients and their families and did not evaluate violence from staff members, which could have similar, if not more, effects on individuals who were exposed to it.

It is alarming that nearly one third of psychiatrists in China have reported an intention to leave their current position. We have also found that WPV, including verbal and physical violence, are significant predictors of turnover intention, while mental health has significant mediating effects on turnover intention. Our findings suggest, while measures to decrease WPV are important, policy makers and hospital administrators need to be aware of the mediating role of mental health on psychiatrists. It may be helpful and cost-effective to promote mental health among the psychiatrists to improve their morale and workforce sustainability.

Data are available from the corresponding authors upon reasonable request.

The studies involving human participants were reviewed and approved by the Ethics Committee of Chaohu Hospital of Anhui Medical University (No. 201903-kyxm-02). The Ethics Committee waived the requirement of written informed consent for participation.

FJ, HL, TL, and YL collected the data set. The original idea for the research was developed by YC, PW, and JZ. PW, YC, and LZ conducted the statistical analyses. YC, PW, LZ, and JZ drafted the paper. YH, NC, YT, and JZ were involved with manuscript preparation and revisions. All authors approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.