The impacts of patient mistreatment on healthcare workers’ role behaviors: a study in Chinese Fangcang shelter hospitals

Customer mistreatment refers to unfair and low-quality interpersonal treatment of service personnel by customers during the service process, including cruel hostility, insult, and other acts of intensity less than physical violence [12]. Walker et al. (2014) found that customer mistreatment as a source of stress can induce negative emotions in employees [13]. Customer mistreatment indirectly affects the organization’s negative reputation through employees’ OCB [14]. Specifically, in the healthcare service industry, customer mistreatment is referred as patient mistreatment. However, there are limited academic studies on the impact of patient mistreatment on healthcare workers, particularly in Fangcang shelter hospitals.

According to conservation of resource theory, stress results from circumstances involving threatened or actual loss of valued resources, and individuals are motivated to defend, conserve, and acquire the resources they value [15]. Being judged or criticized by others (such as leaders, customers or patients) would lead to a loss of resources that are valuable in meeting job demands. Especially for healthcare workers, serving patients is itself a process of consuming physical and psychological resources [16‐18]. Healthcare workers are suffering anxiety due to the heavy workloads in Chinese Fangcang shelter hospitals, which has depleted a great deal of the healthcare professionals' resources to perform their duties [19]. The occurrence of patient mistreatment is not only unable to replenish the original emotional resources of healthcare providers but also requires them to consume additional emotional resources to cope with it [20, 21]. If healthcare professionals’ existing resource reserves are repeatedly depleted without being replenished, it may provoke harmful emotional and psychological reactions, such as emotional exhaustion. In addition, patients' rude and aggressive behavior would impose a greater demand on healthcare professionals who need to spend more resources to regulate their emotions just in order to sustain a normal doctor-patient relationship [22‐24]. For example, the established organizational rules or regulations require healthcare workers to suppress their overt expressions of negative emotions. As a result, healthcare professionals would have to exercise more energy to regulate their emotions, which would lead to an overuse of valuable resources, resulting in emotional exhaustion [25]. Based on the above analyses, the following hypothesis is proposed:

In a service environment characterized by uncertainty and interdependence, employees need to go beyond the narrow scope of work to serve customers [26]. The distinction between in-role and extra-role behavior is reflected in the discretionary and formalization of employee service behaviors [27]. In-role behavior refers to the organization’s members’ activities within the provisions that must be performed to receive the corresponding remuneration [28]. Essentially, it belongs to the employees’ daily work tasks, usually for service performance [29]. The extra-role behavior is the spontaneous initiative of employees, which transcends the role and organizational remuneration system [30], such as autonomous work effort and OCB.

Resources are a set of important factors that can help individuals achieve their goals or meet their psychological needs [31]. Faced with customer mistreatment, employees must consume additional resources to engage in effective self-motivation to maintain good performance [32]. Studies have shown that emotional exhaustion during COVID-19 has a significant negative effect on unhealthy diets of healthcare workers [33]. In the case of emotional exhaustion, healthcare professionals will be more inclined to reduce the input of other resources, such as reducing work input or engagement, to compensate for the resource loss [34]. Due to the loss of resources, workers may spend time and attention calming themselves down, which reduces their involvement and attention at work and leads to a decline in service performance [35]. Healthcare professionals who have lost resources due to the increasing job stress from patient mistreatment, will be more protective of existing resources to prevent further loss. Thus, they may reduce their commitment and responsibility to patients by reducing patient-oriented OCB [36]. In general, patient mistreatment leads to a loss of emotional resources, which may further lower the initiative and motivation of healthcare professionals to provide services, thereby reducing their interests and willingness to serve patients [37]. Accordingly, we propose the following hypotheses:

Dodge et al. (1987) considered reactive aggression as an individual's defensive response to anger when confronted with frustration [38]. DAP is a particular type of reactive aggression. To eliminate anger, the attacked person will try to retaliate against the attacker [39]. However, they are often unable or unwilling to attack the attacker (e.g., their supervisor), then they often turn to attacking other innocent parties [40]. For example, workers show negative emotional and behavioral reactions toward their families after feeling dissatisfied with their leader [41], and studies have also shown that work-family conflicts can lower employees’ emotions and even engage in aggressive behavior toward those around them [42].

Based on the analysis above, we attempted to incorporate displaced aggression as a boundary condition to explore ways to mitigate the negative effects of patient mistreatment. According to self-verification theory, individuals in the process of self-concept formation will continuously receive, integrate, interpret, and modify external information, thereby affecting their self-concept [43, 44]. Self-verification theory emphasizes how healthcare professionals see themselves and what they see from their patients [45]. In order to maintain self consistency, they exhibit behaviors that are consistent with their self-concept [46]. On the one hand, it is the responsibility of healthcare workers to provide high-quality medical services to patients and care for them [47]. Their professional identity is also characterized by kindness and care [48]. Healthcare workers tend to view being mistreated as a failure of service, thereby reducing their self-esteem and self-concept [49]. However, when workers witness patients frequently expressing anger toward themselves, other healthcare providers, and/or shelter hospital managers, they may catch those signals through their empathic abilities. Previous studies have found that healthcare professionals with high empathy skills are more likely to interpret patients' "indiscriminately" attacks as a result of their dilemma, rather than intentionally mistreating others [10]. We predict that when healthcare workers encounter DAP, they tend to regulate their emotions with a high level of care due to seeking consistence between environment and their self-concept. On the other hand, self verification assessments can also reduce emotional exhaustion by confirming a person's self-concept [50]. People hope to reduce uncertainty about their self-concept and ensure that they maximize their enthusiasm for self-awareness [51]. Subsequently, they may construct a personal understanding of the mistreatment and build an interpretation that patients "indiscriminately" attack people around because of anxiety, nervousness, or depression rooted in the treatment of coronavirus disease in Fangcang shelter hospitals. Being mistreated by patients should not be attributed to the workers’ own professional and technical competence, thus reducing the risk of depletion of their self-concept and the consumption of their own resources. Accordingly, we propose the following hypothesis:

Combining H3a, H3b and H4, we propose a moderated mediation model of DAP. If the mediation hypotheses H3a, H3b and the first-stage moderation hypothesis H4 hold, it is reasonable to expect indirect effects of patient mistreatment on role behaviors through emotional exhaustion would also vary according to the different levels of DAP. Therefore, we put forward the following hypotheses:

We summarize our theoretical model in Fig. 1.

The online survey was administered through a questionnaire website (https://​www.​wjx.​cn/​), with the link delivered to qualified participants with the help of medical team leaders and nursing team leaders in Fangcang Shelter Hospital in December 2022. To prevent unnecessary administrative coercion, each questionnaire was anonymized, and each IP address could only be submitted once. The sample consisted of medical and nursing staff from the Yinglong, Yutong, Yuelai, and Nanping shelter hospitals. In total, 639 questionnaires were distributed. After excluding invalid questionnaires that failed to pass the attention check, 493 valid responses were obtained. The demographic characteristics of the participants are presented in Table 1.

A back-translation procedure was conducted to ensure the accuracy and consistency of the translated scales [52]. A bilingual management professor translated items into Chinese from the original English, and then another management professor translated items from Chinese back into English. Discrepancies were checked, discussed, and finalized by both professors. The scales used in this study have been widely applied in academic research. The original scales were modified appropriately to fit the research situation. The specific scales were as follows:

Gender, age, working years, education, technical title, working weeks in shelter hospitals and daily working hours were used as covariates.

Statistical analyses were conducted using SPSS 26.0 and Mplus 8.3, and descriptive statistics were computed for all variables. Considering common method variance (CMV hereafter) that might bias our findings, we introduced a variable that is completely unrelated to this study in theory as a marker variable [58], namely, software self-efficacy [59, 60]. This variable was measured on a 3-item scale. Sample items included “I could complete the job using the software package if there was no one around to tell me what to do as I go.” Cronbach's α for the scale was 0.834. In addition, we tested the hypotheses using hierarchical regression analysis, bootstrap test and conditional process analysis (specifically, moderated mediation in this study).

We used hierarchical regression method to test the main effects [62]. As shown in Table 4, patient mistreatment was positively associated with emotional exhaustion (β = 0.526, p < 0.001), and emotional exhaustion was negatively associated with both service performance (β = -0.105, p < 0.01) and patient-oriented OCB (β = -0.095, p < 0.05). Thus, H1 and H2 are supported.

Referring to the mediation analysis steps proposed by Zhao et al. (2010) [63], Table 5 shows the results of the mediating analysis using 5,000 bootstraps. The indirect effect of patient mistreatment on service performance was -0.055, and the confidence interval was (-0.093, -0.019), which did not include the value of zero, indicating that emotional exhaustion had a mediating effect. Similarly, emotional exhaustion mediated the impact of patient mistreatment on patient-oriented OCB. Thus, both H3a and H3b are supported.

Hierarchical linear regression analysis was performed to test the moderating effect of DAP. Table 6 illustrates the moderating effects of DAP. DAP weakened the impact of patient mistreatment on emotional exhaustion (β = -0.107, p < 0.05).

Figure 2 shows that for high DAP, patient mistreatment has a weaker positive impact on emotional exhaustion. Thus, H4 is supported.

We further tested the moderating mediation effect on the indirect path using conditional process analysis. As shown in Table 7, when service performance was the dependent variable, the bootstrap 95% confidence intervals were (-0.066, -0.017) for high DAP and (-0.109, -0.027) for low DAP, and the difference bootstrap 95% confidence intervals were (0.002, 0.064), indicating the presence of moderated mediating effects. Similarly, the difference in the indirect effects was significant when patient-oriented OCB was the dependent variable. In summary, H5a and H5b are supported.

Finally, to overcome the limitation that moderated mediation can only be tested for discrete moderator values (namely, mean and mean ± one standard deviation) [64], we used the Johnson-Neyman method to depict the continuous confidence intervals of indirect effects. Figures 3 and 4 visualized the moderating mediation effect of different DAP levels (after standardization), indicating that the effect of patient mistreatment on service performance and patient-oriented OCB through emotional exhaustion was weaker when perceived displaced aggression by patients was higher.

This study has several limitations. First, self-reports were used to measure the key constructs. Studies have shown good validity of the self-report measures used in this study [77]. However, the data collection was based on the overall perceptions of healthcare staff in Fangcang shelter hospitals. In the future, changes in healthcare workers’ emotional states could be explored by using a more detailed longitudinal research design or contextual experimental approach. Second, shelter hospitals are relatively narrow and confined environments in which patients tend to "indiscriminately" attack healthcare workers. Frontline healthcare workers who understand the psychological stress and negative emotions of patients are more inclined to regulate their emotions to meet the needs of patients when encountering mistreatment. Future research can be conducted in other contexts to better understand the antecedents and outcomes of patient mistreatment and develop appropriate strategies to address it in different healthcare scenarios.