Figures
Abstract
Introduction
In China, the severity of medical disputes has greatly increased during the past two decades, which has caused various adverse outcomes for health professionals. Previous research on violence in healthcare settings has primarily examined the occurrence of patient-initiated violence and its effects on physicians, but few studies have focused on the impact of the extent of physicians’ exposure to violence. This study examined the different frequency levels of specific types of violence and their relationships to physicians’ psychological wellbeing, including emotional exhaustion (EE), job satisfaction (JS), and intention to leave (IL).
Methods
Using a stratified random sampling method, the participants were drawn from 123 public hospitals in Shanghai, Hubei Province, and Gansu Province in China, and 1,656 completed questionnaires were collected. Chi-square test, analysis of variance, and mixed linear model were employed in the analysis.
Results
The results showed that the rate of exposure to verbal abuse (VA) was the highest (92.75%), followed by threats of assault (TA, 88.10%) and physical assaults (PA, 81.04%). Physicians’ reported high-frequency exposure to VA, TA, and PA was 35.14%, 27.72%, and 19.32%, respectively. The results indicated that exposure to violence significantly affected EE, JS, and IL, and the intensity of the harm of high-frequency exposure was several times stronger than that of low-frequency exposure. Interestingly, we found that VA produced a greater adverse impact on physicians’ EE and satisfaction with work (JS-1) than did TA or PA. This finding may be attributed to the fact that physicians are more likely to be exposed to a high frequency of VA, and the effect of high-frequency exposure is much stronger.
Citation: Shi J, Wang S, Zhou P, Shi L, Zhang Y, Bai F, et al. (2015) The Frequency of Patient-Initiated Violence and Its Psychological Impact on Physicians in China: A Cross-Sectional Study. PLoS ONE 10(6): e0128394. https://doi.org/10.1371/journal.pone.0128394
Academic Editor: Qinghua Sun, The Ohio State University, UNITED STATES
Received: November 8, 2014; Accepted: April 27, 2015; Published: June 1, 2015
Copyright: © 2015 Shi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: All relevant data are available from the Figshare database, under the DOI: http://dx.doi.org/10.6084/m9.figshare.1341888
Funding: Study design, data collection and analysis, preparation of the manuscript and decision to publish: Supported by the National Natural Science Foundation of China (Grant number 71203032), http://www.nsfc.gov.cn/. Preparation of the manuscript and decision to publish: Supported by the Shanghai Planning Project of Philosophy and Social Science (Grant number 2014EGL005), http://www.sh-popss.gov.cn/newsInfo.asp?idval=5869. Preparation of the manuscript and decision to publish: Supported by the Fudan University Scientific Research Capability Promotion for Youth.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Workplace violence refers to violent acts that are directed toward people at work or on duty [1]. It encompasses various types of violence, including verbal abuse (VA), aggression, harassment, bullying, and physical violence such as physical assaults (PA) and threats of assault (TA) [2,3]. Patient-initiated workplace violence in healthcare settings has become an increasingly serious problem in many countries [4–7]. Health professionals in the UK, especially general practitioners (GPs), are consistently at risk of this type of violence [4]. A study on Australian GPs found that 63.7% of those sampled had experienced violence at work [5]. One survey reported that the prevalence of violence toward physicians in emergency departments in Morocco was 70% [8]. In the US, deteriorating physician-patient relationships have been evident since the 1990s [9–11]. In developing countries, a recent study showed that more than half of the surveyed health personnel experienced at least one incident of violence in 2011 [7]. In China, the severity of medical disputes has been increasing significantly over the past two decades. Due to a lack of trust between physicians and patients and an imperfect legal system that provides weak safeguards of the rights and interests of both healthcare professionals and patients, patient-initiated violence against physicians has been on the rise. Patient-initiated violence has become one of the most important issues in China’s new healthcare system reform [12–13].
Workplace violence or aggression can result in a variety of adverse outcomes for health professionals, including poor psychological status, poor work performance, and negative impact on quality of life [8, 14–19]. Empirical evidence has shown that these adverse outcomes lead to an increase in physician turnover intentions and deteriorating relationships between physicians and patients [7, 20–21]. In ethical reviews of research involving human subjects, specific risks are frequently identified, including the probability of a given harm and its severity and duration [22–24]. This assessment approach can also be applied to studying the harm of patient-initiated violence in medical practices. Previous studies have examined the occurrence of patient-initiated violence and its effects on physicians, and the focus tends to be on specific violent incidents. However, few studies have focused on whether different levels (i.e., low probability versus high probability) and different types of violent events (i.e., VA, TA, or PA) produce different impact.
The aim of this study was to investigate the diverse effects of different levels and types of patient-related violence on physicians in terms of emotional exhaustion, job satisfaction and turnover intention. In this study, frequency of violence is defined as the number of times violence occurs within a work setting during a particular period of time. This measure reflects physicians’ perceptions of their work environment, including both directly experienced and indirectly witnessed violent cases. The hypothesis was that the psychological impact of workplace violence varied depending on the level of exposure (i.e., low level vs. high level) and the forms of violence (i.e., VA, TA, or PA).
Methods
Sample and procedures
A cross-sectional survey was conducted using a stratified sampling method in Shanghai, Hubei Province, and Gansu Province in China from June through October 2013. The selection of regions was designed to capture varying levels of socioeconomic status and geographic distributions. We first selected three regions that represented high, middle, and low levels of socioeconomic status that were located in the eastern, central, and western parts of China. Then, we selected three districts or prefecture-level cities/counties representing high, middle, and low levels of socioeconomic status within each province/municipality for a total of nine sampled areas. We randomly selected two tertiary hospitals, three to four secondary hospitals, and six to twelve healthcare community centers/township hospitals in each city/county. In Shanghai, six tertiary general hospitals were selected from the region as a whole (because tertiary general hospitals were distributed very unevenly among the districts). In some counties of Hubei and Gansu provinces, the number of physicians in township hospitals was very small, so we increased the number of sampled township hospitals and decreased the number of surveyed physicians within each township hospital in these counties. Ultimately, the study subjects were recruited from 123 hospitals, including 75 healthcare community centers/township hospitals, 30 secondary hospitals, and 18 tertiary public general hospitals. A total of 2,210 physicians affiliated with these hospitals (with at least one year of work experience at the institution) were invited to participate in this survey, and 1,656 valid questionnaires were collected.
Data collection and measurement
Anonymous, self-administered questionnaires were distributed and collected by coordinators who had been trained to understand the study’s aims, the questionnaire content, and ethical issues. The questionnaire covered the dependent variables, including emotional exhaustion (EE), job satisfaction (JS), and intention to leave (IL). The independent variables consisted of reported patient-initiated violence, participants’ personal characteristics, and other work-related factors.
Hypotheses of relationships between patient-initiated violence and psychological variables.
Previous studies have primarily focused on the occurrence of patient-initiated violence and its effects on physicians. However, few studies have focused on whether different levels (i.e., low versus high frequency) and types of violence may produce different effects. We hypothesized that exposure to different frequencies and types of violence would produce different effects. In model typeI (MT-I), we modeled the relationship between the occurrence of three types of patient-initiated violence and physicians’ EE, JS, and IL. It was hypothesized that the relationships between patient-initiated violence and both EE and IL would be positive, whereas the relationship with JS would be negative. In model typeII (MT-II), we modeled the impact of different levels of violence on the psychological variables. Next, it was hypothesized that compared with a low frequency of violence, associations between high frequency of violence and the dependent variables would be considerably enhanced (see Table 1).
Measurement of emotional exhaustion.
In this study, EE was assessed using the Maslach Burnout Inventory [25], which includes the following 4 items: (1) I feel emotionally drained from my work; (2) I feel exhausted at the end of the work day; (3) I feel fatigued when I get up in the morning and have to face another day on the job; and (4) Working all day is really a strain for me. The response scale included the following 7 categories: never, rarely, occasionally, sometimes, often, frequently, and always. The Cronbach’s alpha of EE (0.934) and the split-half reliability (0.917) were acceptable. With regard to validity, the standardized regression weights of each item on this scale were in the range of 0.863 to 0.909, and the average variance extracted (AVE) [26] of this scale was 0.784. Thus, the tests showed that the convergent validity was good.
Measurement of job satisfaction.
JS was measured using 6 items derived from Konrad et al.’s Global Satisfaction Scales that are specifically for physicians [27], including satisfaction with work (JS-1) and satisfaction with career (JS-2). An example of work satisfaction was: “Overall, I am pleased with my work”. An example of career satisfaction was:“If I were to choose over again, I would still become a physician”. The Cronbach’s alpha coefficients of JS-1 and JS-2 were 0.918 and 0.903, respectively, and the split-half coefficients were 0.898 and 0.870, respectively. Furthermore, the data showed that the convergent validity of the scales was acceptable, with the standardized regression weights of each item for JS-1 (0.834–0.966) and JS-2 (0.784–0.927) in the appropriate range. The AVE for JS-1 and JS-2 was 0.794 and 0.764, respectively. The participants’ responses were recorded on a 6-point Likert scale that ranged from 1 (extremely dissatisfied) to 6 (extremely satisfied).
Measurement of intention to leave.
In this study, IL was designed based on Williams et al.’s research [28], reflecting three types of intention: intention to leave front-line clinical services within one’s hospital (IL-1), intention to leave one’s hospital (IL-2), and intention to quit practicing medicine (IL-3). Three types of IL were identified: (1) I would most likely change my work from front-line clinical services to back-office support in this hospital within the next 2 years; (2) I would most likely leave this hospital within the next 5 years; and (3) I would most likely quit practicing medicine within the next 5 years. The responses ranged from extremely disagree (= 1) to extremely agree (= 6).
Measurement of patient-initiated violence.
Following a previous study [29], patient-initiated violence was measured with questions that asked whether the respondent had directly experienced or indirectly witnessed patient-initiated violence in his/her hospital, including the following items: (1) How often have you or your colleagues suffered from VA by patients or their relatives/friends in the past 12 months? (2) How often have you or your colleagues suffered from TA by patients or their relatives/friends in the past 12 months? (3) How often have you or your colleagues suffered from PA from patients or their relatives/friends in the past 12 months? The response categories were: never, rarely [several times (<12 times) in the past 12 months], occasionally (once per month), sometimes (2–3 times per month), often (once per week), frequently (2–5 times per week), and always (almost every day). The occurrence of a specific type of violence was dichotomized as yes (ranging from rarely to always) or no (never). The frequency of patient-initiated violence was categorized into 3 levels: none (never), low (rarely, occasionally, and sometimes), and high (often, frequently, and always).
Control variables.
The control variables included participants’ personal characteristics (i.e., gender, age, educational level, professional qualifications, years in the profession, annual income, health status, initial vocational motivation, and personal toughness) and work-related factors (i.e., hospital type, work location, department affiliation, night shift, relationships with colleagues and staff, and organizational ethical climate).
The participants’ health status was self-reported on a 5-point scale: very poor, poor, not bad, good, and excellent. Vocational motivation was measured by asking, what was your most important initial reason for becoming a physician: enthusiasm for medicine, solely to make a living, family’s wishes, or other? The personal toughness scale, which was borrowed from Lu and Liang [30], has demonstrated good reliability and validity. In this study, the Cronbach’s alpha coefficient of the personal toughness scale was 0.901. The average of the values of all items was used to reflect personal toughness. The participants’ relationships with colleagues and staff were assessed using the following: how well do I work with my colleagues and with support staff? The responses were recorded on a 6-point scale (“definitely not well” = 1 to “very well” = 6). The organizational ethical climate scale included 5 items that were developed for the current study based on a literature review and consultations with experts. An example is: “For poor patients, my hospital will do its best to treat them and never skimp on treatment or send them away”. The responses were recorded on a 6-point scale (“strongly disagree” = 1 to “strongly agree” = 6). The Cronbach’s alpha coefficient of the organizational ethical climate scale was 0.737. The average of the item values indicated the organizational ethical climate.
Statistical analysis
All of the data were analyzed using SAS Software 9.20 (SAS Institute Inc., Cary, NC, USA). To analyze the psychological status of the respondents, we calculated the average scores for all of the items used to measure EE, JS-1 and JS-2. Basic descriptive statistics was used to analyze personal characteristics, work-related factors, patient-initiated violence, and the psychological variables. Chi-square test was applied to compare the differences in exposure to various levels of frequency of workplace violence among two or more groups. Analysis of variance (ANOVA) was used to explore discrepancies in personal toughness, relationships with colleagues or staff, and organizational ethical climate among different frequency levels of violence. We also estimated multivariate models of the relationship between various types of patient-initiated violence and EE, JS (JS-1, JS-2), and IL (IL1-IL3), respectively. We used a mixed linear model with the function “PROC MIXED” in SAS 9.20 to allow for correlation across respondents within hospitals. The mixed linear model was conducted in two models mentioned in Table 1. Both models were implemented while controlling for personal characteristic variables and work-related factors. The statistical models were separately tested for VA, TA, and PA.
Ethics Statement
Research ethics approval for this study was granted by the Institutional Review Board (IRB) of the School of Public Health, Fudan University (ref: 2012-03-0338) with a waiver of documentation of informed consent for two reasons: (1) this study presented minimal risk of harm to its subjects; and (2) the questionnaires were returned anonymously. During the investigation, each potential subject read a written statement about the survey’s purpose and content and the subject’s rights and interests. All of the subjects provided informed consent.
Results
Distribution of personal characteristics and work-related factors
The respondents’ demographics and work-related factors by types of public healthcare organization in China are shown in Table 2. Overall, the proportions of respondents who practiced in community health centers/township hospitals, secondary hospitals, and tertiary public hospitals were 24.64%, 37.62%, and 37.74%, respectively. Slightly more than half of the physicians were male (54.05%); 47.04% were 30–39 years old; approximately half or more held bachelor’s degrees (60.51%), were residents and below (47.55%), had 1–10 years of experience in the profession (51.39%), had an annual income of less than 50,000 RMB (50.54%), and reported very poor, poor, or not bad health status (65.52%). “Enthusiasm for medicine” and “solely for a living” were cited as the major initial vocational motivation by 45.71% and 33.88% of the respondents, respectively. The average personal toughness score was 4.85 (SD = 0.85), with physicians who worked in tertiary public general hospitals displaying the lowest scores.
Of the 1,656 participants, 32.37% were from Shanghai and 33.94% were from Gansu Province. These physicians came from a variety of hospital departments. Furthermore, 70.95% worked the night shift. Physicians in higher-level hospitals (i.e., tertiary hospitals) were more likely to report poorer working relationships with colleagues and staff and a poorer organizational ethical climate.
Percentages of exposure to the three types of patient-initiated violence at different frequency levels
The percentages of exposure to different frequencies of VA, TA, and PA by personal characteristics and work-related factors are displayed in Table 3. Generally, exposure to VA was the most likely (n = 1536, 92.75%), followed by TA (n = 1459, 88.10%) and PA (n = 1342, 81.04%). Physicians were also most likely to report high-frequency exposure to VA (n = 582, 35.14%), followed by TA (n = 459, 27.72%) and PA (n = 320, 19.32%). Male physicians were more likely to experience high exposure to VA (40.67%), TA (33.52%), and then PA (24.58%). In addition, physicians aged 30–39 reported higher exposure to VA (40.31%) and TA (31.19%). There were no differences between physicians with different professional qualifications, except that residents and below were less likely to report a high frequency of TA (24.74%, p = 0.042). Physicians with a master’s degree or above and those who reported poor or very poor health status were more likely to report a high frequency of VA, TA, and PA. Those who had worked for less than 10 years were the highest-risk group, and those with greater than 20 years of experience were the least likely to report violence. With respect to initial vocational motivation, physicians with an enthusiasm for medicine were at the lowest risk, and those who worked solely for a living were at the highest risk. Physicians with lower scores of personal toughness were more likely to report a high frequency of violence.
Exposure to a high frequency of violence was more likely to occur in higher-level hospitals. Physicians in the departments of internal medicine, surgery, and gynecology/obstetrics/pediatrics were exposed to a higher frequency of violence than those in other departments. Physicians in surgery departments were most likely to report a high frequency of VA, TA, and PA (44.01%, 38.32%, and 29.34%, respectively). A greater proportion of physicians who worked night shifts reported a high frequency of violence. In addition, physicians who were exposed to a high frequency of violence were more likely to work in hospitals with lower average scores for organizational ethical climate and had poorer relationships with their colleagues and staff.
Mixed linear model analysis
Table 4 describes the associations between types of violence and physicians’ EE, JS, and IL. The fixed effects of the mixed linear model analysis clearly demonstrated that violence had significant adverse effects on EE, JS, and IL. Interestingly, of the three types of violence, exposure to VA caused the largest adverse impact on EE (β = 0.811, p<0.001) and satisfaction with work (JS-1) (β = -0.255, p<0.01), and the coefficients of VA were the largest in MT-I. With regard to career satisfaction (JS-2), only PA produced a significant effect (β = -0.274, p<0.001). With regard to IL, the results differed among the three types of IL. MT-I shows that both VA and TA increased physicians' intentions to leave front-line clinical services (βVA = 0.244, p< 0.05; βTA = 0.270, p<0.01), but these types of violence exerted no significant impact on physicians’ intention to leave the hospital or intention to quit practicing medicine. In addition, among the three types of IL, only physicians' intention to quit practicing medicine was affected by PA (β = 0.198, p<0.05).
However, when physicians were separated into low-frequency and high-frequency groups, the results changed (as shown in MT-II). The intensity of harm resulting from high-frequency exposure was several times stronger than that of low-frequency exposure in relation to EE. With regard to JS-1, compared with MT-I, MT-II showed that only high-frequency exposure to VA (β = -0.494, p<0.001) and TA (β = -0.369, p<0.001) significantly influenced this variable. With regard to JS-2, MT-II showed that exposure to a high frequency of VA and TA also generated a considerably adverse impact. Both low and high frequencies of PA had remarkable effects on two types of JS. With respect to IL, for all three types of intention to leave, low-frequency exposure to VA, TA, or PA had little impact, whereas high-frequency exposure was related to significantly greater IL. Exposure to a high frequency of PA had the greatest impact on intention to leave front-line clinical services (β = 0.653, p<0.001), intention to leave the hospital (β = 0.687, p<0.001), and intention to quit practicing medicine (β = 0.700, p<0.001). Additionally, exposure to a high frequency of VA (β = 0.488, p<0.001) or TA (β = 0.607, p<0.001) exerted the largest impact on intention to leave front-line clinical services, whereas exposure to a high frequency of PA (β = 0.700, p<0.001) had the greatest influence on physicians' intentions to quit practicing medicine.
In the mixed linear model analysis, all of the models were calculated while controlling for the personal characteristics variables and work-related factors.
Discussion
The results showed that more than 90% of the respondents were exposed to at least one type of violence in the 12 months prior to the survey. The occurrence rates of VA, TA, and PA were all considerably high (92.75%, 88.10%, and 81.04%, respectively). Moreover, the proportions of physicians who reported exposure to a high frequency of VA, TA, and PA were 35.14%, 27.72%, and 19.32%, respectively. Of the violence types, the risk of VA was the most serious. The deterioration of the relationship between physicians and patients in China has been noted [13,31–32]. In this study, physicians’ exposure to violence was significantly related to their higher EE, lower JS, and higher IL. Notably, we found that the adverse impact of high-frequency exposure to violence on physicians was greater than that of low-frequency exposure. Effective measures must be taken to end violence against physicians in China.
The potential predictors of exposure to patient-initiated violence
Previous studies have indicated several potential predictors of exposure to patient-initiated violence, such as gender, age, educational level, years in the profession, and department [5,18,33], but their conclusions have not been consistent. One study that used Australian urban data reported that violence was more likely toward female GPs and that more experienced GPs encountered less violence [5]. A study that used data from Fujian and Henan Provinces in China showed that male medical professionals, those who were middle aged (30–45 years), and those with a bachelor’s degree experienced a higher incidence of violence [18]. A study of GPs in the UK found that sex was only a predictor of sexual harassment and that number of years in the profession was a predictor only of intimidation [33]. In this study, male physicians, those with a master’s degree or above and those who had 1–10 years of experience in the profession were more likely to experience a high frequency of violence. These differences may be partially due to cultural factors, methodological differences such as the definitions of various types of violence, or sampling techniques.
However, we posit that female physicians may have better communication skills and, thus, may be better able to manage some potentially problematic situations. Physicians with more years in the profession may be more experienced in medical and social skills. It seems counterintuitive that physicians with a higher educational level are more likely to be exposed to a high frequency of violence. In China, most young physicians have a master’s degree or higher, especially in tertiary hospitals. Patients typically assume that physicians with a higher educational degree should have better medical capabilities and skills. If the outcomes do not meet a patient’s expectations, disputes may occur.
Additionally, physicians with poorer health status, lower personal toughness, who worked in tertiary hospitals located in Shanghai or Hubei Province, who specialized in surgery, and who worked the night shift were more likely to be exposed to a high frequency of violence. Tertiary hospitals are more likely to treat patients with more difficult diseases. Medical accidents that occur in surgical departments may produce more significant damage to patients and their families than those that occur in other departments; thus, conflicts may be more likely to occur, especially TA or PA. It has been reported that violence is more likely to occur between 4 p.m. and 8 a.m. [34]. Other researchers have supported this finding [1, 20]. The higher rates of violence during the evening and night hours may be due to the types and conditions of patients who seek treatment during later hours, such as intoxicated patients and those with violence-related injuries [35].
More attention to physicians’ exposure to a high frequency of violence
Previous studies have explored the associations between workplace violence and healthcare professionals’ psychological status, such as job stress, anxiety, and job satisfaction [21,36–40]. However, it was not clear from these studies whether the intensity of the adverse effects changed with the frequency of violence. The current study found that exposure to VA, TA, or PA was significantly related to higher EE and lower JS-1, but the intensity of these adverse effects multiplied for those in the high-exposure group. Consequently, low-frequency exposure to VA or TA does not significantly decrease JS-2, but high-frequency exposure notably decreases JS-2. Furthermore, high-frequency exposure to VA, TA, or PA was related to greater IL. The effect of high-frequency exposure to PA on IL was much greater than that of exposure to VA and TA. This result is similar to that of another study that found that bullying is a significant determinant in predicting intention to leave [39].
People generally believe that PA is more dangerous than VA. As a result, the adverse effects of VA are often ignored. The current study showed that VA produced a greater adverse impact on physicians’ psychological status than did TA or PA for EE and satisfaction with work (JS-1). We suspect that physicians are more likely to be exposed to a high frequency of VA, and the effect of high-frequency exposure is much stronger. Physicians may be able to tolerate low-frequency exposure to violence. However, an increase in the frequency of violence can generate a qualitative change; for example, the rate of turnover intention increases significantly, and more physicians feel disappointed with their careers. This result is supported by a study’s [31] findings that more physicians have abandoned medicine and that significantly fewer physicians want their children to work in the medical profession in China than in the past.
Strengths and limitations
The major strengths of this study include its reasonably large sample size and the empirical finding that high-frequency exposure to VA, TA, and PA has a greater effect on physicians’ EE, JS, and IL than does low-frequency exposure. Furthermore, VA has a greater effect than the other types of violence in the healthcare setting. However, this study also has various limitations. First, the data were based on self-reports, which may involve recall/report bias. Second, causal associations between violence and the physicians’ psychological variables cannot be established in this study because of its cross-sectional nature. It will be necessary to conduct a prospective study to identify causal relationships. Third, the duration of exposure to violence was not collected in this study because it was difficult for the respondents to recall the timing of such events in a retrospective survey. Furthermore, it is worth noting that the measurement of patient-initiated violence in our study covered both experienced and witnessed cases because we intended to explore the effect of such exposure in a work environment with a low or high frequency of violence. Thus, the rate of occurrence of violence that we measured was for a work setting and not for individual physicians. Therefore, the reported occurrence of violence was relatively high. Finally, although our sample size was reasonably large, the results may not be generalizable to the whole nation because the study took place in only three regions of a very large country.
Conclusion
The results reveal a serious problem in China’s healthcare system and indicate that reducing the frequency of violent incidents and creating a safer work environment for physicians should be a top priority for both government and society. All effective measures should be taken without delay. Hospital administrators, healthcare policy makers, and society should pay more attention to physicians’ high exposure to violence.
Acknowledgments
We gratefully acknowledge the significant contributions of the following people in gathering information for this study: Xuefeng Wei, Hongbo Zhu, Huiqing Tang, Jianjun Gu, Wenxiu Li, and Qilin Jin. The authors also appreciate all of the hospitals that provided assistance with the data collection for this research project. Additionally, the authors would also like to thank the anonymous reviewers for their insightful comments and suggestions.
Author Contributions
Conceived and designed the experiments: PZ SW DX XKZ. Performed the experiments: PZ JWS YZ FB. Analyzed the data: PZ JWS SW. Contributed reagents/materials/analysis tools: PZ LYS. Wrote the paper: PZ JWS SW.
References
- 1. AbuAlRub RF, Khalifa MF, Habbib MB. Workplace violence among Iraqi hospital nurses.J Nurs Scholarsh. 2007; 39: 281–288. pmid:17760803
- 2. Ceramidas DM, Parker R. A response to patient-initiated aggression in general practice: Australian professional medical organisations face a challenge. Aust J Prim Health. 2010; 16: 252–259. pmid:20815996
- 3. Tolhurst H, Baker L, Murray G, Bell P, Sutton A, Dean S. Rural general practitioner experience of work-related violence in Australia. Aust J Rural Health. 2003; 11: 231–236. pmid:14641220
- 4. Elston MA, Gabe J, Denney D, Lee R, Beirne MO. Violence against doctors: a medical (ised) problem? The case of National Health Service general practitioners. Sociol Health Illn. 2002; 24:575–598.
- 5. Magin PJ, Adams J, Sibbritt DW, Joy E, Ireland MC. Experiences of occupational violence in Australian urban general practice: a cross sectional study of GPs. Med J Aust. 2005; 183: 352–356. pmid:16201952
- 6. Kowalenko T, Walters BL, Khare RK, Compton S. Workplace violence: a survey of emergency physicians in the state of Michigan. Ann Emerg Med. 2005; 46:142–147. pmid:16046943
- 7.
World Health Organization. New research shows workplace violence threatens health services. 2002 May 10 [cited 8 November 2014]. In World Health Organization. Available: http://www.who.int/mediacentre/news/releases/release37/en/.2012.
- 8. Belayachi J, Berrechid K, Amlaiky F, Zekraoui A, Abouqal R. Violence toward physicians in emergency departments of Morocco: prevalence, predictive factors, and psychological impact. J Occup Med Toxicol. 2010; 5:27. Available: http://www.occup-med.com/content/5/1/27. pmid:20920159
- 9. Emanuel EJ, Dubler NN. Preserving the physician-patient relationship in the era of managed care. JAMA. 1995; 273(4): 323–329. pmid:7815662
- 10. Warren MG, Weitz R, Kulis S. Physician satisfaction in a changing health care environment: the impact of challenges to professional autonomy, authority, and dominance. J Health Soc Behav. 1998; 39: 356–367. pmid:9919857
- 11. Arnetz JE, Aranyos D, Ager J, Upfal MJ. Development and application of a population-based system for workplace violence surveillance in hospitals. Am J Ind Med. 2011; 54: 925–934. pmid:21739469
- 12. Wang X, Zheng J. How to end violence against doctors in China. Lancet. 2012; 380: 647–648. pmid:22901881
- 13. Ending violence against doctors in China. Lancet. 2012; 379: 1764. pmid:22579308
- 14. Kivimäki M, Virtanen M, Vartia M, Elovainio M, Vahtera J, Keltikangas-Järvinen L. Workplace bullying and the risk of cardiovascular disease and depression. Occup Environ Med. 2003; 60: 779–783. pmid:14504368
- 15. Carmi-Iluz T, Peleg R, Freud T, Shvartzman P. Verbal and physical violence towards hospital- and community-based physicians in the Negev: an observational study. BMC Health Serv Res. 2005; 5: 54. Available: http://www.biomedcentral.com/1472-6963/5/54. pmid:16102174
- 16. Coles J, Koritsas S, Boyle M, Stanley J. GPs, violence and work performance—‘just part of the job?’. Aust Fam Physician. 2007; 36: 189–191. pmid:17339990
- 17. Shields M, Wilkins K. Factors related to on-the-job abuse of nurses by patients. Health Rep. 2009; 20: 7–19. pmid:19813435
- 18. Wu S, Lin S, Li H, Chai W, Zhang Q, Wu Y, et al. A study on workplace violence and its effect on quality of life among medical professionals in China. Arch Environ Occup Health. 2014; 69: 81–88. pmid:24205959
- 19. Kadra G, Dean K, Hotopf M, Hatch SL. Investigating exposure to violence and mental health in a diverse urban community sample: data from the South East London Community Health (SELCoH) Survey. Plos One. 2014; 9(4): e93660. pmid:24691206
- 20. Gates DM, Ross CS, McQueen L. Violence against emergency department workers. J Emerg Med. 2006; 31: 331–337. pmid:16982376
- 21. Heponiemi T, Kouvonen A, Virtanen M, Vänskä J, Elovainio M. The prospective effects of workplace violence on physicians’ job satisfaction and turnover intentions: the buffering effect of job control. BMC Health Serv Res. 2014; 14, 19. Available: http://www.biomedcentral.com/1472-6963/14/19. pmid:24438449
- 22.
National Bioethics Advisory Commission (NBAC). Defining, interpreting, and applying concepts of risk and benefit in clinical research involving children. In: Marilym J, Field , Behrman RE. The ethical conduct of clinical research involving children. Washington DC: The National Academies Press; 1998. pp.114–145.
- 23. Rid A, Emanuel EJ, Wendler D. Evaluating the risks of clinical research. JAMA. 2010; 304:1472–1479. pmid:20924013
- 24. Westra AE, Wit JM, Sukhai RN, De Beaufort ID. How best to define the concept of minimal risk. J Pediatr. 2011; 159: 496–500. pmid:21764402
- 25. Maslach C, Jackson SE. The measurement of experienced burnout. J Organ Behav. 1981; 2: 99–113.
- 26. Fornell C, Larcker DF. Evaluating structural equation models with unobservable variables and measurement error. J Mark Res. 1981; 48:39–50.
- 27. Konrad TR, Williams ES, Linzer M, McMurray J, Pathman DE, Gerrity M, et al. Measuring physician job satisfaction in a changing workplace and a challenging environment. Med Care. 1999; 37: 1174–1182. pmid:10549620
- 28. Williams ES, Konrad TR, Scheckler WE, Pathman DE, Linzer M, McMurray JE, et al. Understanding physicians’ intentions to withdraw from practice: the role of job satisfaction, job stress, mental and physical health. Adv Health Care Manag. 2001; 2: 243–262.
- 29. Wynne R, Clarkin N, Cox T, Griffiths A. Guidance on the prevention of violence at work. The European Commission, Directorate general V; 1997.
- 30. Lu GH, Liang BY. Development of hardiness scale. Studies of Psychology and Behavior (Chinese journal). 2008; 6:103–106.
- 31. Li J. New generations of Chinese doctors face crisis. Lancet. 2012; 379: 1878. pmid:22608337
- 32. Gong Y, Han T, Chen W, Dib HH, Yang G, Zhuang R, et al. Prevalence of anxiety and depressive symptoms and related risk factors among physicians in China: a cross-sectional study. Plos One. 2014; 9(7):e103242. pmid:25050618
- 33. Koritsas S, Coles J, Boyle M, Stanley J. Prevalence and predictors of occupational violence and aggression towards GPs: a cross-sectional study. Br J Gen Pract. 2007; 57: 967–970. pmid:18252072
- 34. Ergün FS, Karadakovan A. Violence towards nursing staff in emergency departments in one Turkish city. Int Nurs Rev. 2005; 52: 154–160. pmid:15842328
- 35. McAneney CM, Shaw KN. Violence in the pediatric emergency department. Ann Emerg Med. 1994; 23: 1248–1251. pmid:8198298
- 36. Driscoll RJ, Worthington KA, Hurrell JJ. Workplace assault: an emerging job stressor. Consult Psychological Practical Research. 1995; 47: 205–212.
- 37. Quine L. Workplace bullying in nurses. J Health Psychol. 2001; 6: 73–84. pmid:22049239
- 38. Sherman MF, Gershon RR, Samar SM, Pearson JM, Canton AN, Damsky MR. Safety factors predictive of job satisfaction and job retention among home healthcare aides. J Occup Environ Med. 2008; 50: 1430–1441. pmid:19092499
- 39. Simons S. Workplace bullying experienced by Massachusetts registered nurses and the relationship to intention to leave the organization. ANS Adv Nurs Sci. 2008; 31: E48–59. pmid:18497581
- 40. Belayachi J, Berrechid K, Amlaiky F, Zekraoui A, Abouqal R. Violence toward physicians in emergency departments of Morocco: prevalence, predictive factors, and psychological impact. J Occup Med Toxicol. 2010; 5, 27. pmid:20920159