The healthcare industry has been identified as one of the most violent workplace sectors1, with the emergency departments (ED) ranked as one of the areas with the highest risk for such violence.2Patient-related aggression and violence have been identified as common behaviors in the modern ED.3Within this context, nurses have been identified as the occupation with the most risk of patient-related violence, with 60-90% of nurses reporting exposure to verbal and physical violence.2Emergency department nurses have been identified as having the most stressful workplace setting of all nurses4, and are exposed to a disproportionate amount of violence.5High levels of verbal abuse and threatening behavior are reported6,with up to 90% of ED nurses having experienced physical violence at some point in their careers and all have reported experiencing verbal abuse.7
The term violence encompasses a wide range of behaviors8, from physical assault or direct violence to non-physical forms of violence such as verbal abuse and sexual harassment.9Studies have reported that up to 20% of nurses have experienced intimidation, harassment or assault of a sexual nature.3Evidence suggests that the term “violence” is frequently used by nurses as a broad term to describe everything from witnessing verbal abuse through to being a victim of physical assault.10
Patients are consistently identified in the literature as being the most common source of violence against nurses8; responsible for up to 89% of cases.11Other sources of violence include friends and relatives accompanying patients and horizontal violence from other staff members.3
Verbal abuse has been identified as the most common form of abuse experienced by nurses12, and has been labelled a global phenomenon.13It includes rudeness, shouting, sarcasm, swearing, unjustified criticism, ridicule in front of others, threat of personal harm to the person, their family or property, rumor mongering12, as well as sexual innuendo.14Of these behaviors, swearing has been identified as the most common form of verbal abuse14, and includes both face-to-face and telephone abuse.15
Physical violence is frequently preceded by verbal violence and it may constitute a warning sign for potential or impending physical violence.2It has been defined as any intentional physical contact, actual or threatened, and may or may not result in an injury to the victim.16It can also include overt behavior designed to intimidate or threaten, for example, punching a wall or throwing furniture.6The ready availability of and easy access to hospital equipment has resulted in objects such as scissors, syringes, needles and stretcher poles being used as weapons against nurses.13
Violence and assault against nurses is reported to be on the increase15; however it is widely acknowledged that the true level is unknown due to chronic under-reporting of violent incidents17in Australia and internationally.3Estimations of under-reporting range from 20% up to 90%8,16,18,19, and are referred to as the “dark figure” of workplace violence.12A culture of silence is said to exist8, meaning that accurate statistical analysis of the incidence and prevalence of violence has become impossible.3Reasons cited for failure to report a violent incident include insufficient time, a feeling of resignation that no benefit will come from the process, lack of appropriate support and feedback, the belief that such violence is part of the job8; fear of reprimand and lack of knowledge about reporting procedures.14The overall findings indicate that nurses feel unsupported by management in relation to workplace violence.15
Nurses in EDs in particular have been identified as significantly under-reporting violent incidents; a fact thought to be related to the high levels of violence endemic in these areas.5These high levels of violence have resulted in a desensitization on the part of many nurses to the point where violence has become an expected and accepted part of their job.20Verbal and physical abuse are regarded as occupational hazards11, and there is a rationalization on the part of many nurses that such violence is unavoidable.21
The types of injuries sustained by nurses as a direct consequence of violent behavior from patients range from minor, for example scratches, to major including fractures and loss of consciousness11, up to extreme cases where nurses have been stabbed and even killed.22Australian figures for the period 2000 to 2002 reported 3621 incidents involving patients and physical violence or violent verbal exchanges against health care staff and in 5% of these incidents staff were injured.20
Even in the absence of physical injury, nurses have been found to experience moderate to severe psychological reactions for up to 12 months following an episode of violence.9Exposure to violence has been linked to long-term psychological effects, including Post-traumatic Stress Disorder and burnout.23Despite the seriousness of the consequences, informal discussion or debriefing with colleagues was the most common coping mechanism referred to in the literature.24
Emotional effects reported in the literature include guilt, self-doubt, feelings of professional incompetence25, anger, powerlessness, unhappiness, degradation, shame, fear, astonishment, antipathy towards the perpetrator24, and sleeplessness.15Nurses have reported feeling more cautious and deriving less satisfaction from their patient-related care, as well as being fearful at work. This fear and caution can result in low morale and lead to a situation where patients are avoided.25Studies have identified a negative correlation between violence experienced by health care staff and patient-related quality of care.25
The health system is faced with increased costs in terms of sick leave, decreased productivity, staff turnover and attrition and workers compensation payments due to injuries resulting from patient-related violence.15Quantifying the cost of patient-related violence is difficult as it includes intangible outcomes such as loss of morale, difficulties with retention and recruitment of staff, impact on patient care and therapeutic relationships and negative public relations which are difficult to measure in monetary terms.26
For the purpose of this review, the following definitions of terms were used:
Violence: a range of behaviors from non-physical behaviors and verbal abuse through to overt acts of physical violence.
Patient-related: patients presenting to the ED, including the parents of pediatric patients. Whilst groups other than patients were not the focus of the review, for example, relatives, some incidental results have been included.
Nurses: all classifications of nurses were considered in this review.
Emergency department: the department in a hospital that is staffed and equipped to provide rapid and varied emergency care, especially for those who are have suffered trauma or sudden and acute illness. For the purposes of this review, the term “emergency department” was used, however it can be used interchangeably with “accident and emergency” and" emergency Service".
The objectives of this review are to identify, appraise and synthesize the best available evidence on the impact of patient-related violence against nurses working in EDs.
Types of participants
The quantitative component of this review will consider studies that include ED nurses of all classifications.
The qualitative component of this review will consider studies that include ED nurses of all classifications.
Types of intervention(s)/phenomena of interest
The quantitative component of the review will consider studies that evaluate ED nurses' experiences with patient-related violence.
The qualitative component of this review will consider studies that investigate ED nurses' experiences with patient-related violence.
Types of studies
The qualitative component of the review will consider studies that focus on qualitative data including, but not limited to, designs such as phenomenology, grounded theory, ethnography, action research and feminist research.
Only primary research was included for the purpose of this review. In the absence of research studies, other texts such as opinion papers and reports were not considered. Original research was considered as this is subject to a rigorous process that ensures the validity and reliability of the results.
Types of outcomes
This review will consider studies that include the following outcome measures:
• Frequency of episodes of violence
• The types of violence - physical and verbal
• Organizational reporting of episodes of violence
• Antecedents and risk factors; Impact and sequelae of violence: psychological and physical effects
• Prevention and control measures to manage episodes of violence.
The search strategy aims to find both published and unpublished studies. A three-step search strategy will be utilized in this review. An initial limited search of MEDLINE and CINAHL will be undertaken, followed by analysis of the text words contained in the title and abstract and of the index terms used to describe the article. A second search using all identified keywords and index terms will then be undertaken across all included databases. Thirdly, the reference lists of all identified reports and articles will be searched for additional studies. Studies published in English will be considered for inclusion in this review. Studies published between January 2001 and December 2011 will be considered for inclusion in this review. The most current evidence on the topic will be sourced; therefore the inclusion criteria included only studies from the last 10 years.
The databases to be searched include:
- Mosby's Index
- Cochrane Central Register of Controlled Trials (CENTRAL).
- Google Scholar
- The following keywords will be used in the search strategy:
- Emergency Department or ED or Emergency Service
- Accident and Emergency or A & E or Triage
- Violence or Aggression or Assault or Abuse
- Patient or Patient-Related or Client
Method of the review
Quantitative papers selected for retrieval will be assessed by two independent reviewers for methodological validity prior to inclusion in the review using standardized critical appraisal instruments from the Joanna Briggs Institute Meta-Analysis of Statistics Assessment and Review Instrument (JBI-MAStARI) (Appendix III). Any disagreements that arise between the reviewers will be resolved through discussion, or with a third reviewer.
Qualitative papers selected for retrieval will be assessed by two independent reviewers for methodological validity prior to inclusion in the review using standardized critical appraisal instruments from the Joanna Briggs Institute Qualitative Assessment and Review Instrument (JBI-QARI) (Appendix III). Any disagreements that arise between the reviewers will be resolved through discussion, or with a third reviewer.
Quantitative data will be extracted from papers included in the review using the standardized data extraction tool from JBI-MAStARI (Appendix IV). The data extracted will include specific details about the interventions, populations, study methods and outcomes of significance to the review question and specific objectives.
Qualitative data will be extracted from papers included in the review using the standardized data extraction tool from JBI-QARI (Appendix IV). The data extracted will include specific details about the interventions, populations, study methods and outcomes of significance to the review question and specific objectives.
Qualitative research findings will, where possible, be pooled using JBI-QARI. This will involve the aggregation or synthesis of findings to generate a set of statements that represent that aggregation, through assembling the findings rated according to their quality and categorizing these findings on the basis of similarity in meaning. These categories are then subjected to a meta-synthesis in order to produce a single comprehensive set of synthesized findings that can be used as a basis for evidence-based practice. Where textual pooling is not possible, the findings will be presented in narrative form.
Description of studies
The initial search strategy identified 10,297 papers, of which 55 were deemed potentially relevant to this review and selected for retrieval, based on assessment of their titles and abstracts. After removal of duplicates and detailed examination, full text papers were retrieved for 33 of these studies. One paper was identified through hand searching of the reference lists of these papers. After analysis of methodological quality, 20 papers were included for data extraction and analysis of results. Details of the search strategy are provided in Appendix I. Details of the studies excluded from the review following methodological assessment with reasons for exclusion are provided in Appendix VI. The details of the selection process and search results are presented in Appendix II. The 18 papers remaining were included in this review.
Description of included studies
Appendix V provides details of the included studies that reported the impact of patient-related violence on nurses working in EDs in narrative and summary table form. A total of 18 papers were included in the review, consisting of 11 quantitative, four qualitative and three mixed method studies.
The studies were drawn from six countries as detailed in Table 1 below.
Of the 11 quantitative studies that utilized a questionnaire, five provided details of the validity and reliability of their instruments, such as Cronbach's alpha or test-re-test scores.14,27-30The remaining papers provided such details on some aspects of their survey tool31,32, or did not provide any details.33-36
Factors and Outcomes reported
The main measures reported in the quantitative studies included:
• Frequency of episodes of violence
• The types of violence - physical and verbal
• Antecedents and risk factors - including perpetrators of violence
• Impact and sequelae of violence: including personal and professional effects of violence on ED nurses such as flashbacks, sleeplessness, nightmares, depression, fearfulness, time taken off work and severity of violence.
The qualitative papers reported measures such as:
• Impact and sequelae of violence : including injuries and lost workdays due to physical assaults
• Antecedents and risk factors
• Prevention and control measures to manage episodes of violence, including coping mechanisms
• Organizational reporting of episodes of violence.
The mixed-methods studies reported their data using the following measures:
• Antecedents and risk factors - including perpetrators of violence and predictive behaviors
• The types of violence - physical and verbal
• Impact and sequelae of violence, including personal and professional effects.
Eight studies were appraised using QARI, of which seven papers were included in this review and one paper excluded. Twelve studies were appraised using MASTARI, with 11 papers included in this review and one excluded.
A summary of each study is presented in narrative form below.
Quantitative studies on violence in the ED
Violence against nurses in hospitals: prevalence and effects30
The authors conducted a cross-sectional study to determine the degree and effects of violence among ED nurses (n = 81) at a Kuwaiti hospital. Factors and outcomes related to violence were evaluated using three questionnaires developed by the researchers. The authors used a 12-item frequency-weighted questionnaire that they had developed and validated in an earlier study to measure rates, frequency and severity of violence. The researchers used data from the Income Data Services Study to develop a five-item, duration-weighted questionnaire that was used to measure the effects of violence at work.37The items in the instrument measured: reliving the experience (flashbacks), sleeplessness, depression, fearfulness and time taken off work. The last instrument assessed the views of nurses on violence including training, hospital policies regarding reporting and management of offenders following aggressive incidents. The reliability of the questionnaire was tested using Cronbach's alpha, which yielded reliability coefficients, based on six items, of alpha = 0.8730 and standardized item alpha = 0.8728.
A total of 81 of the targeted 86 ED nurses completed the questionnaire, a response rate of 94%; however the authors do not provide any demographic details of their sample. Of these, 70 (86%) reported having experienced some form of violent episode with 719 incidents noted, with a mean of 3.3 (standard deviation 1.2); of these 679 were classified as mild and 40 as moderate. The authors used a list of 12 responses which were subsequently categorized as mild, moderate or severe violence. It should be noted that this was not a conclusive list and opened up the possibility of types of episodes being missed, for example, “threatened to hit” and “threatened with a knife/gun” are included, but no mention is made of other types of threats such as threats to career, to harm family, to destroy property, or to rape nurses. All nurses surveyed reported experiencing verbal abuse and 13 (19%) reported moderate acts of violence, defined by the authors as single acts of physical violence unlikely to result in serious injury. None of the nurses reported severe levels of violence, defined by the authors as being attacked with a weapon that was likely to have resulted in a serious or fatal injury.
The majority of the participants (n = 71, 96%) reported that they had suffered from after-effects lasting for up to four weeks after experiencing violence, with three nurses experiencing fearfulness for more than 28 days. A statistically significant number of male nurses experienced more incidents of violence than females and reported that they suffered from fearfulness (p < 0.01 and p <0.05 respectively); however it should be noted that only a small number of males (n = 16) were included in the sample of 81.
Although only 19% of the nurses had received training to deal with potentially violent patients, 44% believed that training to deal with potentially violent incidents in the ED would have been useful. Of the 40 episodes of moderate violence involving physical attacks, only five of the offenders were charged by the police.
Violence towards health care workers in emergency departments in west Turkey33
The author employed a cross-sectional design and used a survey to collect data about episodes of workplace violence with ED staff. The authors claim that a random sample was selected; however they do not provide any details of how this process took place. A total of 195 staff out of a potential population of 242, from 18 private, university or public health institutions in Turkey responded to the survey: a response rate of 81%. The ED staff included nurses (n = 73) and other healthcare professionals such as physicians, ward aides and admitting clerks. No details of the survey were provided and no exact numbers of specific professional groups or demographic information were reported.
For the purpose of analysis, the researchers defined and divided types of violence into three categories: verbal/emotional abuse, specific threats and physical action which included overt acts such as slapping and kicking. Participants were asked to recall experiences of such violence directed at them during the course of work over the previous year.
Statistical differences were analyzed using the Pearson's chi-squared statistical test where appropriate, with the level of significance set at 5%. Results were reported as means and standard deviations for continuous data.
Of the 195 respondents who completed the survey, 72% (n = 141) had experienced some form of violence in the preceding 12 months. Of these, 70% (n = 98) had experienced verbal/emotional abuse; 53% (n = 75) specific threats and 9% (n = 12) physical action. Participants reported that the perpetrators of violence included patient's family members or relatives (n = 125, 89%) and patients themselves (n = 73, 52%). Analysis by age groups revealed that participants in the 30-39 year age group experienced the highest proportion of any type of violence (82.1%, p < 0.05). Nurses experienced the highest proportion of the different forms of violence (81%), followed by physicians (78%); and almost 56% of nurses reported experiencing specific threats. Analysis by profession indicated that a higher proportion of nurses (42.4%) and those suturing/dressing wounds (46%) reported violence from patients (p < 0.05). Although none of the participants reported major injuries such as lacerations or fractures, half of those who experienced physical violence received support such as attending the emergency service and taking days off. Although a higher proportion of women than men reported experiencing some form of violence, with the exception of specific threats, differences in the gender and status of emergency health staff were not statistically significant (p > 0.05). This result has cultural significance with the authors postulating that this may be because violence towards women is more prevalent due to male dominance in this society.
Violence towards emergency department nurses by patients14
The authors used a descriptive, longitudinal cohort study design to determine the incidence of patient-related violence towards nurses in two EDs in Australia. A sample of 71 nurses was recruited from a population of 108, with a response rate of 66%. Four instruments were developed to collect data and included a demographic details form, a brief violence record, data extraction form and a violence questionnaire. All instruments were tested in a pilot study to ascertain their reliability and face validity. The demographic details form and Violence Questionnaire were adapted from an earlier study by Murray and Snyder.38Face validity of the Violence Questionnaire was assessed by a pilot study with 10 ED nurses and test-retest reliability was established by determining stability over time and the percentage of agreement between the testers with a result of 91%, indicating good consistency. Inter-rater reliability was established for the Data Extraction Form by using Pearson's correlation (r = 0.96), which was statistically significant p<0.001. For the patient's triage category inter-rater reliability of nursing triage was established using Kendall's tau-b (0.86), which was statistically significant (p = 0.01).
A brief violence record was completed by participants after each episode of violence and these forms were collected weekly by the researchers. Those nurses who had been subjected to violence were then issued with a detailed violence questionnaire. Hospital records were used to determine the total number of presentations to the ED in the five months of the study.
Fifty nurses (70%) reported 110 episodes of violence in a five-month period: approximately five per week, or two per 1000 presentations. Patients were identified as the perpetrators in all of these episodes. Patient-specific factors identified included alcohol intoxication (n = 30, 27%), substance abuse (n = 27, 25%) and behaviors associated with mental illness (n = 42, 38%). In addition, the nurses reported that patients often displayed demanding behaviors and requested attention prior to exhibiting violent behavior (n = 48, 44%).
Verbal violence was the most common type of violence experienced with a total of 58 (53%) episodes of verbal violence and 29 (26%) of both verbal and physical violence reported. The most common types of verbal violence experienced were swearing (n = 57, 61%) and yelling (n = 40, 36%), while pushing (n = 11, 10%), hitting (n = 3, 3%), and kicking (n = 6, 3%) were the most commonly experienced physical behaviors. Triage was the highest risk area for all types of violence and this violence occurred most often on evening shifts (n = 41, 37%).
Workplace violence against Iranian nurses working in emergency departments27
The authors used a cross-sectional study to sample 196 nurses from 11 EDs in teaching hospitals in Tehran, Iran, to determine the frequency and nature of physical and verbal workplace violence against these nurses using a recall period of one year. The authors reported a response rate of 95% (n = 186). They adapted a self-administered questionnaire, previously developed by the International Council of Nurses, World health Organization and the Public Services International. This survey was previously validated by the Joint Program on Workplace Violence in the Health Sector.39While no definition of violence was mentioned in the article, a review of the original tool does contain a definition from the World Health Organization; however it is unclear whether the authors adopted this. This survey tool was translated into Persian, and content and construct validity assessed by five nursing faculty members before being piloted for clarity, ease of administration and applicability with 20 nurses. The authors decided to omit questions regarding sexual and racial harassment due to the scarcity of their occurrence in Iranian EDs; however, no evidence was given to support this decision.
Data were analyzed using only descriptive and inferential statistics. The chi-square test was used to determine the association between nurses' gender and violent incidents. While the authors report a statistically significant p-value of <0.05, it should be noted that only a small number of male nurses participated in the study, with the majority of the sample (89%) identified as female.
Specific demographic details were not provided; however, the sample was described as being largely young female nurses with less than five years nursing experience. Thus the characteristics of the sample may limit the generalizability of the results. The authors reported that the majority of nurses (92%) had experienced verbal abuse in the preceding year, while 20% had experienced physical abuse. Patients' relatives were the most common source of such violence. Under-reporting of episodes of violence was identified, with 85% (n = 152) of participants reporting that no reporting procedures were available.
Violence against nurses working in US emergency departments31
The authors conducted a cross-sectional study to investigate ED nurses' experiences and perceptions of violence from patients and visitors in EDs in the United States. A convenience sample of 3465 was recruited from a population of approximately 31,905 members of the US Emergency Nurses Association (ENA), which corresponded to a response rate of approximately 11%. The researchers developed a 69-item online survey that collected data about nurses' personal experiences with physical violence and verbal abuse in the ED in the preceding three years and the policies and procedures of the respondents' hospital and ED for addressing workplace violence. The survey also included items about the respondents' beliefs about the precipitating factors of violence and barriers to reporting violence in the ED. The online survey was developed using Survey Select Expert (version 5.6) by an Emergency Nurses' Association work team and the authors state that the survey was evaluated by experts for content validity; however they do not provide details of this. It was then pilot tested on a sample of 15 ED nurses.
The chi-squared test of association and Fisher exact test were used to compare independent groups with respect to percentages and the Kruskal-Wallis and Mann-Whitney-Wilcoxon tests were used to compare independent groups with respect to non-categorical variables. For all statistical analyses a significance level of 0.05 was used and data presented as a mean with standard deviation.
Results indicated more than 50% of respondents had experienced physical violence which included being spat on, hit, pushed/shoved, scratched and kicked. Verbal abuse such as being yelled/cursed at, intimidated and harassed with sexual language/innuendo was reported by 70% of the respondents. One third of participants reported having considered leaving their ED or emergency nursing because of violence.
Twenty-three percent (n = 811) of respondents had experienced a high frequency of physical violence (> 20 times) from patients/visitors in the ED during the past three years. About 20% (n = 604) of respondents reported experiencing a high frequency of verbal abuse (> 200 times) from patients/visitors during the same period. Nurses who worked night shifts and weekends were more likely to experience frequent physical and verbal violence. Participants identified a number of barriers to reporting violent episodes which included: the perception that reporting episodes might have a negative effect on customer service; ambiguous reporting policies; fear of retaliation from management, hospital administration, nursing staff, or physicians; the perception that reporting episodes of violence was a sign of incompetence or weakness; lack of physical injury to staff; the attitude that violence comes with the job; and a lack of support from administration/ management. These barriers were associated with an increased risk of experiencing violence in the ED. Nurses who felt that there were no barriers to reporting ED violent incidents were much less likely to have experienced frequent ED physical violence than were other nurses: 15.4% versus 28.5% (p<0.001).
The main factors reported as precipitants of ED violence included: alcohol intoxication; substance misuse; care of psychiatric patients in the ED; crowding/high patient volume and prolonged wait times.
Violence against nurses and its impact on stress and productivity32
The authors undertook a cross-sectional study to collect data on workplace violence experienced by ED nurses in the United States. A survey was sent to a randomly selected sample of 3000 members of the Emergency Nurses' Association and 230 participants returned the survey: a response rate of only 9%. The survey was composed of four separate sections. The first section asked the participants to describe in narrative a single recent episode of workplace violence that caused them the most stress. The second section consisted of 22 Likert-type items based on the Impact of Events Scale-Revised.40This set of items evaluated the presence and magnitude of post-traumatic stress symptoms during the seven days after a traumatic event; however it should be noted that in this study no limit was set on when this event initially occurred, which creates the potential for recall bias and inconsistencies between results due to variations in time elapsed since the event occurred. The authors reported that the Impact of Events Scale-Revised had high internal consistency ratings (0.79-0.91) and strong sensitivity (74.5) and specificity (63.1).
The third section consisted of a 29-item Healthcare Productivity Survey. The instrument was developed and validated by the researchers in a previous study to measure the perceived change in work productivity after exposure to a stressful event.41Internal consistency reliability and test-retest reliability of the scale was reported as (0.871 - 0.945) and (r = 0.801, p < 0.001) respectively, with a sample of emergency nurses.41The last section included demographic questions regarding their age, gender, race, education, care population, the geographic location of their ED and whether their employer provided violence prevention training or critical incident stress debriefing.
The results confirmed that workplace violence was a significant problem for those ED nurses who participated in the study and was associated with experiences of distress, decreased work productivity, and reduced quality of nursing practice. Thirty-seven percent (n=82) of the respondents had a negative total productivity score on the Healthcare Productivity Survey demonstrating decreased performance after a violent event. The mean score on the Impact of Event Scale-Revised group was 18.67 (range 0-83), indicating presence of at least one stress symptom after a violent event. There was a correlation reported between the total Healthcare Productivity Survey and Impact of Event Scale-Revised scores; however this was not statistically significant (p = 0.07).
The prevalence of study participants with post-traumatic stress symptoms during the seven days was considered clinically significant, with 17% of participants having scores high enough for a probable diagnosis of Post-Traumatic Stress Disorder and 15% with scores associated with suppressed immune system functioning.
Violence against emergency department workers35
The authors used a cross-sectional design to measure the violence experienced by ED workers from patients and visitors in five hospitals in the United States. A recall period of six months was used in an attempt to reduce the potential for recall bias. A sample of 242 ED workers was recruited from a target population of approximately 600 workers, which resulted in a response rate of 40%. Nurses comprised the largest professional group with 95 participants (39%), followed by physicians (n = 49, 20%), patient care assistants (n = 27, 11%) and another six from professional and ancillary groups of staff. A survey consisting of a combination of multiple choice, open-ended and Likert-type items was developed by the authors to collect data from participants; however the survey tool was not validated.
The term “violence” was defined by the authors and for the purposes of the survey, classified as: verbal harassment, sexual harassment, verbal threats or physical assaults. Data about demographics, previous violence prevention education and frequency of physical and non-physical violent acts in the preceding six months were collected. Other survey items measured injuries and lost workdays due to physical assaults, frequency of reporting assaults and variables related to the assaults. Likert-type items were used to measure participants' feelings of safety and levels of satisfaction with their job, the ED, the hospital and security.
Data were analyzed using descriptive statistics. Verbal harassment by patients was reported by 98% of the nurses surveyed (n = 93), which was the highest result for hospital staff. Incidents of verbal threats from patients were highest for physicians (n = 41, 83%), followed by nurses (n = 74, 78%), though nurses experienced the greatest percentage of verbal threats from visitors (67%). Forty-four percent of nurses reported experiencing at least one incident of sexual harassment from patients (n = 42).
During the study period, there were 319 episodes of physical violence initiated by patients and at least 10 by visitors. Nurses were the professional group most frequently targeted, experiencing 67% of the 319 episodes of physical violence. Those who worked in the psychiatric ED reported the highest frequency of being assaulted: seven or more times by a patient (11%). Nurses (8%) were also recipients of the most episodes of physical violence by visitors. Thirty-two injuries were reported during the study period including bruises, bites, abrasions and scratches. Of the 115 participants who reported experiencing at least one assault by a patient, 65% had never reported the incident in their organization.
Antecedents reported included patient and visitor factors, such as alcohol and drug use, psychiatric diseases, dementia and an inability to deal with a crisis situation. Staff factors included lack of adequate staff, lack of knowledge about a patient's previous history of violence and lack of violence prevention training. Hospital and environmental factors perceived to be significant included long waiting times for patients, lack of security or police presence and environmental design of the department.
The authors suggested that it was imperative to increase efforts aimed at decreasing the incidence of violence and to ensure that any form of violence is neither accepted nor tolerated in any healthcare setting. The study was based on self-reported data; however the authors surveyed the previous six-month period to reduce the risk of recall bias associated with this type of data collection. In addition, the anonymous nature of the survey meant that there was no way to compare the characteristics of respondents with those of non-respondents.
A survey of workplace violence across 65 US emergency departments28
The authors analyzed data from a nationwide survey across 65 sites in the USA. Data were obtained from the National Emergency Department Safety Study (NEDSS) which examined clinical processes and systemic factors contributing to patient safety in hospital EDs across a five-year period. A sample of 3518 health professionals was recruited from a population of 5695; resulting in a response rate of 62%. Of these, 1837 were nurses (52% of respondents), 1395 medical officers, 88 physician's assistants and 169 participants were classified as “other” staff. To develop the survey the researchers revised a previously developed instrument. To further refine the survey, data were obtained from in-depth personal interviews with key informants and focus groups conducted across three EDs, including cognitive testing. Psychometric testing was then conducted at 10 EDs and data from these sites were used to determine the final set of questions prior to its distribution to participants.42
Outcome measures identified included the number of physical attacks against staff; the frequency of guns and knives in the ED and staff perceptions of safety. Multivariate analyses were conducted to determine which respondent and ED characteristics were associated with the perception of safety and which ED characteristics were associated with increased frequency of attacks and weapons. Linear and logistic regression modelling was also employed to assess relationships between outcome measures and independent variables from the survey.
A total of 3461 physical attacks were reported over the five-year period, with a mean of 11 attacks per ED. Nurses were five times less likely to feel safe “most of the time” or “always” compared to the other professional groups (odds ratio OR = 0.21; 95%, CI = (0.16) to (0.28), p<0.0001). Nurses who had worked for more than five years felt less safe than those who had worked shorter periods of time. The authors report that the frequency of weapons, number of attacks, presence of metal detectors and violence training were not statistically significant predictors of staff feeling safe most of the time or always.
Less than 15% of the EDs had metal detectors and EDs with metal detectors reported a higher number of physical attacks and were more likely to have weapons brought to the ED on a daily basis (odds ratio = 26.3, 95%, CI = (2.0) to (339), p=0.001). However this could be attributed to the fact that more weapons were detectable due to the presence of the metal detectors rather than the fact that more weapons were present. Less than half of the EDs had some type of training program for staff and staff who worked in EDs where such training was available, reported a higher perception of safety than those without it; however this finding was of borderline statistical significance (p<0.06).
The authors report that the majority of EDs that participated in the study were in large, academic hospitals that may have different violence rates than other EDs. In addition the authors' state that their sample may not be representative of the United States, with one-third of EDs withdrawing from the study and EDs in the Northwest of the country over-represented. The presence of confounding factors was acknowledged, for example, the definition of what constituted an attack against staff was not standardized and therefore different interpretations may have occurred. Similarly reporting mechanisms were not standardized and the presence of metal detectors was not uniform and may have caused the number of weapons found or recovered to vary. The perceptions of feeling safe at work may have been a reflection of factors not controlled for in the study, for example, staffing levels and organizational response after episodes of violence.
Violence in accident and emergency (A&E): the role of training and self-efficacy36
Lee employed a cross sectional design using a questionnaire to survey a sample of 76 ED nurses from a population of 130 at two study sites, with a response rate of 58% about their experiences in the preceding three months. No details were included about validation of the survey instrument. In addition the study employed The Difficult Behavior Self-Efficacy Scale (DBSES) and respondents were asked to rate their self-efficacy in dealing with aggression.43This scale was previously validated by the authors and confirmed in testing by the researchers; Cronbach's alpha yielded an internal consistency measure of 0.88, indicating good internal consistency of scale items.
Factors and outcomes measured were types of violence; the frequency of violent episodes and the type and amount of aggression management training. No definition of the term “violence” was provided to participants; however verbal and physical violence were differentiated in the results. Self-efficacy was defined as an individual's belief that they can succeed at a given task or behavior.
Ninety-six percent of the nurses (n = 73) had experienced verbal abuse and 79% (n = 60) physical violence in the preceding three months. The mean level of self-efficacy perceived by the participants was average; however there is no definition of the score provided to explain this. The relationships between categorical and interval demographic variables and self-efficacy in managing violent behavior were analyzed using independent t-tests. Higher levels of self-efficacy were associated with having experienced higher levels of verbal aggression in the study period (t = 2.77, df = 74, p<0.01) and being a manager (t = 3.08, df = 69, p<0.01).
The authors described aggression management training regimes as sporadic and fragmented with a lack of consistency between trainers and an absence of any type of refresher program; this was in spite of the mandatory nature of the training. Only two nurses specifically described being taught verbal de-escalation techniques.
Verbal and physical violence in emergency departments: a survey of nurses in Istanbul, Turkey29
This cross-sectional Turkish study used a questionnaire to survey ED nurses about their experiences with violence in the previous 12 months. Two-hundred-and-fifty-five nurses were recruited out of a population of 262: a response rate of 97%. The survey tool was developed by the researchers and tested on an expert panel of nurses and their suggestions incorporated into the final version to facilitate face validity. In addition to this survey the authors used the Attitudes Towards Patient Physical Assaults Questionnaire devised by Poster and Ryan and previously validated by them.44
Outcome measures for the study included the frequency of episodes; the source of violence and days of sick leave taken as a consequence of violence. Descriptive results were reported including percentages and raw values. Ninety-one percent of nurses surveyed (n = 233) had experienced verbal abuse and 75% (n = 191) physical violence, with a range of one to five episodes in the preceding year. The perpetrators were most commonly patients' relatives and friends, responsible for 32% (n = 74) of verbal violence and 62% (n = 119) of physical violence. Patients' relatives and friends, and patients acting together were responsible for 63% (n = 146) of verbal violence and 23% (n = 44) of physical violence. Patients acting alone were identified as the perpetrators in 6% (n = 13) of episodes of verbal violence and 15% (n = 28) of physical violence. Family and friends involved in episodes of violence were typically males, responsible for 77% (n = 57) of verbal abuse and 79% (n = 94) of physical abuse and this was postulated by the authors to be related to the dominant role played by males in Turkish society.
Under-reporting was common with 80% of respondents choosing not to report episodes of verbal and physical violence. Reasons cited included dissatisfaction with responses given by administration to episodes of violence, such as no noticeable follow-up conducted; apathy; fear of losing their job; fear of being blamed and fear of legal procedures that would follow such a report. Those nurses who did report were dissatisfied with the handling of the episode and in 36% of cases of verbal abuse and 42% of physical violence no response was given to the nurse.
The authors report that while over half of nurses who had been exposed to violence felt anxiety about being the subject of another violence incident, the rate of those taking sick leave was only 3%. The authors postulate that this could be because of traditional attitudes and cultural values. The authors reported that in Turkey unequal gender-power relations associated with discrimination mean that women are especially vulnerable to violence. In Turkey, nursing is a profession traditionally practised by women and this was reflected by the participants in this study who were all female. Violence against women, whether nurses or not, is generally treated as a private matter. Thus, violence of all types against women remains largely unreported. The legal system is another obstacle for the victims of violence. According to Turkish law, for legal proceedings to commence, a report must be made to hospital police and then to the local police center. The nurses must then submit to a full body examination and blood sampling to determine if there are any alcohol or drug abuse issues. The brief is then delivered to a public prosecutor who attempts to mediate the dispute prior to going to court. The case is often closed if a perpetrator denies the charges and even where cases are successful, punishment is weak or non-existent. Only one case of physical violence resulted in legal proceedings being initiated against the perpetrator.
Support was received from physicians (33%) and colleagues (40%) in the ED rather than through more formal mechanisms such as administration. Only 3% of nurses took sick leave as a result of an episode of violence; this was consistent for verbal (n = 7) and physical abuse (n = 6). Once a nurse takes sick leave legal proceedings are initiated and so the authors report that many nurses do not take sick leave for this reason. The majority of nurses (n = 211, 83%) had not undertaken any training to assist them in prevention and management of episodes of violence. Only 19 nurses (7.5%) felt safe most of the time while working in the ED. Sixty-five percent of respondents (n = 166) stated that they never felt safe in the ED while 28% (n = 70) felt safe only some of the time. Ninety percent of nurses expected patients and their relatives to exhibit violence towards staff.
Violence towards nursing staff in emergency departments in one Turkish city34
The authors used a descriptive cross sectional study to determine the incidence and attitudes towards violence faced by nurses in the ED in four major hospitals in Turkey. Data were collected using a 34-item questionnaire developed by the authors that collected socio-demographic data and information about their experiences of violence from patients. No details on the validity of the survey were provided by the authors. The recall period used was not defined and results were discussed in terms of “violence experienced throughout the career in the ED”,34(p.156)which creates issues in terms of recall bias and the ability to compare individual results. A definition for the term “violence” was not provided; however results were discussed in terms of verbal and physical violence.
A sample of 66 was recruited from a population of 92 ED nurses: a response rate of 72%. Analyses were conducted using ANOVA, Post Hoc Tukey and Student's t-tests where appropriate. Findings were accepted as statistically significant where the p-value was <0.05. Statistical analysis of the results indicated that nearly all of the respondents had experienced verbal violence (n = 65, 99%). Over half of the participants reported experiencing episodes of verbal violence more than 15 times in their professional careers (n = 35, 54%). Most of these episodes were reported to have occurred in the last three months (n = 52, 80%) and the authors go on to say this “probably meant that this kind of violence occurred frequently in the ED setting”34(p156); however, they provide no justification for this statement. Physical violence was experienced by 20% of the respondents (n = 13).
Most episodes of violence occurred during the night shift and patients' relatives or friends were the most common perpetrators of both verbal (n = 42, 65%) and physical violence (n = 11, 85%), rather than the patients themselves. There was a significant relationship between the age and years of experience of nurses and the frequency of violence. A Post Hoc Tukey test revealed that as age and years of experience increased, the relative number of episodes of verbal and physical violence also increased.
Continuing education concerning the prevention and management of violent behavior in the ED setting was considered to be necessary by 91% of the nurses. Taking legal action after experiencing episodes of violence was considered to be ethical by 91% of the participants.
Qualitative papers on violence in the ED
A descriptive study of the perceptions of workplace violence and safety strategies of nurses working in level one trauma centers45
Catlett conducted a descriptive study looking at the phenomenon of workplace violence. A convenience sample of eight nurses who had experienced violence while on duty was recruited from two level one trauma centers in the United States and semi-structured interviews used to collect data. For the purposes of the study, violence was defined as “the victimization of an RN practicing in a hospital setting by another person or persons characterized by fear, physiological or psychological hardship, or loss”.45(p.520)A demographic profile form and an interview guide designed by the author were used to collect data during the interviews and had been developed during a pilot project with five nurses; however these tools had not been validated. The approach to analysis was based on four main components, namely: bracketing or ethical enquiry, analyzing (empirical examination), intuiting (personal insight) and describing (esthetics). Verbatim transcripts of the interviews were used to identify themes emanating from the collected data. The researcher also kept a journal to minimize bias during the data collection phase. No other methods to enhance rigor were discussed by the author and analysis was not verified by an independent researcher. Member checking was not used and the small sample size means that the results may not be transferable. As with all self-reported data, there is a risk of bias and a possibility that some nurses may have been reluctant to share all aspects of their experiences and outside factors may have affected the interview process.
Two major themes emerged from the data: insufficient safety measures and vulnerability. Issues related to safety measures that were raised by the participants included easy access to the ED with weapons, despite the presence of metal detectors and the lack of training on managing violent situations in the ED. Nurses reported feeling unsafe and vulnerable to violent episodes at work and the strongest expressions of vulnerability were in relation to psychiatric patients. The participants reported that episodes of verbal aggression were mostly experienced from family members, whilst patients with psychiatric disorders or under the influence of alcohol were most likely to display physical violence.
Violence in the emergency department: a culture care perspective46
The authors utilized a phenomenological approach to explore nurses' perceptions regarding workplace violence in the ED. Purposeful sampling was used to recruit a sample of 12 ED nurses; four male and eight female who had experienced some form of workplace violence from a metropolitan trauma center in Midwest US. Semi-structured interviews were used to gather data and data saturation is reported to have occurred after 12 interviews.
Transcripts were analyzed using Leininger's four phases of analysis for qualitative data.47However there is incongruence between the use of phenomenology and the incorporation of findings into an analysis framework (culture care) and the methodology appears to be more qualitative and descriptive in nature. There is minimal description of the process of data analysis and a definition of violence was not provided to participants. After the analysis; member checking was conducted with key participants for their input and clarification; however no changes were required as a result of this process. The authors do not specify how many participants were involved in this process.
The authors identified four subcultures involved with the phenomenon of workplace violence, namely: ED nurses; administration; clients with violent behaviors, and clients without violent behaviors.
Three themes were identified: hospital policies not conducive to supporting a safe ED environment; ED staff not valued as being as important to the hospital administration as the hospital's public image; and anxiety, fear and negative emotions due to violence in their work environment having an impact on the quality of the participants' nursing practice. A lack of trust of the administration culture was evident, with participants perceiving that there was a lack of follow-up following reported episodes of violence. Under-reporting was acknowledged, with most of the episodes of violence described during the interviews not documented, especially in the case of verbal abuse. Barriers cited included negative attitudes and little or no follow-up action from the administration.
The nurses perceived violence to be a daily occurrence. A distinction was made between verbal abuse, for example, being sworn at, called names and threatened, and physical assault, although the terms were not defined. Violent behavior was perceived to be related to cognitive and physical conditions, for example, from patients who were under the influence of drugs and/or alcohol, or those with psychiatric conditions. Socio-economic factors such as geographic origin were also perceived to be significant.
Most of the participants perceived that hospital policies were not conducive to supporting a safe ED environment. Factors specific to the ED perceived to have contributed to episodes of violence included long waiting times, easy access to the department, slow response times from security and a lack of metal detectors. As a protective response, nurses reported becoming increasingly vigilant around patients and visitors and discussed concerns regarding the quality of nursing practice offered as a result of anxiety, fear and negative emotions associated with violence.
Using action research to plan a violence prevention program for emergency departments48
Gates et al. employed action research to investigate employees' and managers' perceptions and experiences regarding violence prevention strategies. This was the first stage in a program designed to plan, implement and evaluate a violence prevention and management intervention. Twelve focus groups (four per study site), were conducted with a total of 96 participants which corresponded to a response rate of 31%. Two focus groups were made up of ED employees and one each for managers and patients. The ED employees were allocated to the groups by stratified random sampling from nurses, physicians, patient care technicians, paramedics, security personnel and radiology technicians. Patients were approached at 15-minute intervals in the ED and asked to participate in a focus group the following weekend. In total, 24 management participants, 47 employees and 25 patients took part in the focus groups which were conducted over a three-month period and ranged in duration from 60 to 75 minutes.
The researchers used the Haddon matrix to develop focus group questions aimed at gathering data about the pre-assault, assault and post-assault time frames and to compare these findings to planned strategies.49This matrix is widely used in public health to understand and identify measures to prevent injury.49
The audio recordings from the focus groups were transcribed and then verified by a research team member. Thematic analysis following Wolcott's method50was undertaken by two researchers independently to identify ideas and patterns, and then used to contextualize these patterns into an existing framework, in this case the Haddon matrix. The results were then compared for congruency and inconsistencies resolved by returning to the original transcript. Finally a third researcher conducted a final review of the transcripts and analysis. The results were presented as before, during and after an assault.
Employees and managers from the different occupational groups agreed that multiple interventions such as examining hospital policies, improving staff education and training and increasing communication with patients and among ED staff were required to prevent and manage violence against ED workers. All groups supported the need for improved communication as a strategy to prevent violence. Employees and managers were also of the view that security and/or police presence in the ED should be increased and that zero tolerance policies needed to be enforced. All participants supported limiting access to the ED as a way to prevent violence.
Employees indicated that existing policies and education about conflict resolution and aggression management techniques were inadequate. They identified the need for frequent educational opportunities aimed at specific strategies to de-escalate situations with patients and resolve conflicts. Additional workplace violence education and training was perceived to be important and participants stated that this should be conducted in an interdisciplinary manner involving all ED employees, security personnel and management. Patients highlighted the need for improved staff communication and comfort measures.
Management and employee participants supported the use of debriefing following episodes of violence; however they reported that this was rarely done. Barriers cited included time constraints, a lack of administrative support and a workplace culture that tolerates violence as part of the job. While reporting mechanisms were acknowledged, barriers to reporting were acknowledged including confusion about which episodes to report, time constraints and a lack of feedback from management and administration about the reported event. Employees stated that reporting should be easy to access and quick to complete and requested that a staff liaison role be created to address any ongoing concerns of the victims of violence.
Patient-related violence at triage: a qualitative descriptive study17
The authors used a qualitative descriptive methodology to uncover the experiences of a group of Australian triage nurses with patient-related violence. A recall period of one month was chosen to limit the possibility of recall bias. A convenience sample of six triage nurses was recruited and the small size and the fact that only one study site was chosen mean that the findings of this study have limited generalizability. Data were collected through semi-structured interviews.
Data were analyzed using qualitative content analysis.51This involved data re-presentation where a descriptive summary of the main points are organized in a way that best fits the data. Member-checking was used to enhance rigor in the study. The study's outcome measures included antecedents and risk factors; types of abuse; the impact of patient-related violence on nurses; and risk management strategies, reporting and coping mechanisms utilized by nurses in dealing with the consequences of such episodes. A definition of violence was provided to participants that included both verbal and physical behaviors.
The authors identified antecedents including patient-specific risk factors such as those with a mental health diagnosis; those under the influence of alcohol and/or illicit substances; younger adults, aged between 16 and 25 years; the parents of pediatric patients and patients from a lower socio-economic bracket. Emergency department specific factors included those secondary to patient volume in the department, for example, long waiting times and time of day; the afternoon shift, especially on the weekend and during winter was identified as a high risk time. Factors specific to nurses included less experienced staff being at high risk and staff attitudes, for example, those who employed a condescending attitude were thought to be at greater risk of experiencing an episode of violence. The authors reported that some participants were able to recognize cues and warning signs in potentially violent patients and use this knowledge to prevent escalation to violence.
Episodes of verbal and physical violence were reported. Swearing was the most common form of verbal abuse reported by five of the six participants, with threats, shouting, making unreasonable demands and intimidation also noted. Physical abuse included nurses being slapped, kicked and hit as well as the use of conventional weapons (knife) and the opportunistic use of hospital equipment as weapons, for example, the contents of a sharps bin.
The impact of patient-related violence on nurses was often mitigated by nurses making a judgment regarding the intent of the episode. For example, unintentional violence from patients judged not to have control over their actions, such as dementia patients were viewed more favorably than intentional violence.
The impact on nurses included feelings of frustration, powerlessness and degradation. Participants reported a diminished empathy for patients and were fearful for their safety. Overall, a sense of the inevitability of patient-related violence was noted and nurses perceived patient-related violence to be increasing in both frequency and severity. Risk management strategies employed by the hospital were perceived to be reactive in nature rather than preventative, and included the use of security and duress alarms and the design of the workplace, for example, the use of safety glass at triage. Training in aggression minimization techniques was incomplete in nature or totally absent. The under-reporting of episodes of violence was common due to time constraints, the non-user friendly nature of the reporting system and the high frequency of episodes. The authors reported that informal debriefing with colleagues was the main coping mechanism in dealing with the aftermath of violence and whilst formal debriefing was desired, it was not available in this workplace.
Mixed-methods studies about violence in emergency department s
Violence against healthcare workers in a pediatric emergency department41
The authors utilized a multiple case study approach to describe work place violence perpetrated by patients and visitors against nurses, physicians and allied health workers in a large urban pediatric teaching hospital in the United States. A purposeful sample of 31 ED workers was recruited from a population of nearly 200 healthcare workers; a response rate of approximately 16%. Of these, 12 (39%) were nurses, eight (26%) physicians and 11 (35%) other allied health personnel. Data were collected using a combination of semi-structured interviews, non-participant observation, digital photographs and archival records. The latter included policy documents and guidelines as well as educational material provided by the hospital. The interviews focused on participants' worst experiences with violence in the preceding six months. Forty hours of non-participant observations were conducted on workers, patients and visitors in the ED, in four 10-hour blocks. Digital photographs were taken of physical artefacts during these observation periods and included the physical environment where episodes of violence occurred; as well as objects identified as important during the observation by researchers and during the interviews by participants. The outcome measures of interest were the perpetrators of violence; types of violence; precipitants; and the personal and professional effects on ED staff.
Transcripts were analyzed using a modified version of the constant comparison method as described by Lincoln and Guba.52This involved a three-stage approach that was conducted independently by two of the authors. Data were categorized and any disagreements between the researchers resolved by discussion. Finally, method triangulation of the data was conducted with comparison to the field notes and archival data. In addition, debriefing and an audit trail were performed to enhance rigor in the study.
The authors reported that in cases of verbal abuse, family members were the most likely source (82%), whilst episodes of physical violence were more likely to be perpetrated by pediatric patients (76%). A mental health diagnosis was reported in 14 of the 16 episodes of patient-initiated violence reported. Verbal abuse included yelling, cursing and threatening workers with physical harm, as well as invading a worker's personal space and the use of symbolic violence, for example, pointing a finger at a worker or blocking their exit.
Precipitants or antecedents included factors specific to workers, patients and the ED itself. Less experienced workers were thought to be at an increased risk of experiencing episodes of violence, while patients with a mental health diagnosis were identified as significant, especially scheduled patients not permitted to leave the department. High risk family members were identified as those with a lack of respect for women or persons in positions of authority; those under the influence of alcohol or illicit substances and those who perceived that the needs of the patient were not being sufficiently addressed. Issues of significance specific to the ED included access to the department, the presence of multiple visitors per patient and issues related to volume in the department, for example, long waiting times, overcrowding in the waiting room, noise levels and perpetrators being stared at in public. Overall the belief was expressed that violence in the community was increasing, leading to a greater acceptance of violence as a normal way to express anger.
The effects of this violence on workers included emotions such as fear, anger and frustration. Workers reported avoidance of perpetrators and a decrease in productivity and the ability to focus on work. They were also concerned with a perceived poor image of the hospital by patients and visitors.
Innocent or culpable? Meanings that emergency department nurses ascribe to individual acts of violence53
Luck, Jackson, and Usher explored the meanings that ED nurses ascribed to acts of violence from patients, their family and friends and what impact these meanings had on how they responded to such acts. The study was conducted in a regional Australian ED on a convenience sample of 20 ED nurses, which represented a response rate of 37%. This was a two-phase sequential mixed method study. Phase 1 involved 50 hours of unstructured participant observation, unstructured open-ended interviews with three nurses and researcher journaling. These data were thematically analyzed and the results used to inform development of Phase II of the study. Phase II involved 290 hours of participant observation, 16 semi-structured open-ended interviews, 13 unstructured interviews, organizational documents and a researcher journal. Quantitative data of violent events were generated using a structured observational guide. The authors report that textual data were analyzed thematically and numeric data were analyzed using frequency counts. Rigor was established by the use of an audit trail by the researchers and member checking of emergent themes and ideas was undertaken in the field interviews.
Sixteen episodes of violence were observed in the 290 hours of participant observation and included in the analysis; however none of these episodes were reported by the nurses involved. The authors assert that the nurses made judgments about specific episodes and these informed and guided their verbal and behavioral response, their decision to report and their short and long-term emotional response.
The study found that episodes perceived to be “within normal limits” on a continuum of violence were not reported by nurses. ED presentations were perceived by the participants to be appropriate or inappropriate and so-called legitimate presentations were afforded greater empathy and tolerance in terms of their behavior. These judgments were based on the degree of personalization of the violence, the presence of mitigating factors and the reason for the ED presentation. Mitigating factors were identified as co-morbid or presenting health problems, including mental health issues, alcohol intoxication, withdrawal, delirium, dementia and head injuries; and psycho-social issues that decreased a perpetrator's ability to react in a rational and informed manner. These could lead to anxiety, confusion and disorientation and/or a lack of appreciation or understanding regarding resources and the processes of an ED, any of which could precipitate acts of violence.
STAMP: components of observable behavior that indicate potential for patient violence in emergency departments54
In another study, Luck et al. utilized the data from their earlier study involving 16 observed violent events to identify components of observable behavior that could indicate a person's potential for violence. Five behaviors were grouped together under the acronym STAMP: staring, tone, agitation, mumbling and pacing, and a cumulative relationship was noted between them. The authors noted that where nurses were able to recognize these signs, they were able to utilize de-escalation techniques.
Staring was identified as an important early indicator and includes both glaring, staring and looking intently as a means of intimidation or the lack or absence of eye contact. The avoidance of eye contact was associated with a passive form of resistance that could escalate into actual physical violence. This was noted in nine of the 16 encounters. The tone and volume of voice was identified in 13 of the 16 episodes and emerged as an important cue both independently and in conjunction with other cues. It included yelling or raising of the voice, urgency in speech, sarcastic or caustic comments, sharp retorts and dismissing or demeaning the nurse through vocal inflections.
Anxiety was observed in patients and recognized by physical symptoms such as looking flushed, hyperventilating and speaking rapidly. It was attributed by the authors to one or more of three broad factors and as anxiety escalated nurses were seen to intervene. Psycho-social, situational and contextual stressors included a loss of a sense of control of patients and elements specific to the ED such as noise and overcrowding. Patients' diagnoses and co-morbidities were identified as exacerbating feelings of anxiety compounding the stressors described previously and included those that led to symptoms of confusion and disorientation, for example, head injuries, mental health issues, delirium, dementia, intoxication and withdrawal from alcohol and hyper or hypoglycemia. Thirteen of the 16 episodes involved disorientated patients who were not able to be nursed in seclusion. Anxiety was also associated with a lack of understanding of the ED process and distribution of ED resources, for example the triage scale.
Mumbling was noted in 11 of the 16 episodes and included comments made just loudly enough to be heard. It also included slurring and incoherent speech; however this was difficult to distinguish from alcohol intoxication in some cases. It was perceived by the authors to be a sign of mounting frustration and involved the asking of repetitive questions, and negative and aggressive statements about the service being received, especially with regards to waiting times. Agitation was most frequently manifested in pacing and also included staggering, flailing or swinging of the arms, resisting interventions by pulling away from the nurse or rejecting therapeutic interventions.
This systematic review was undertaken to investigate the best available evidence on the impact of patient-related violence on nurses working in the ED and provides a summary of the included studies. A systematic search of the literature resulted in 18 published studies that met the inclusion criteria for this review. Of these, 11 were cross-sectional studies that involved the use of a survey tool, four were qualitative studies that employed semi-structured interviews or used focus groups to collect their data and three were mixed-methods studies that used multiple data collection methods.
The lack of homogeneity between the quantitative studies precluded the use of meta-analysis, aspects of which are discussed below.
Context and perpetrators: There were differences in the identification and discussion of the perpetrators of violence between the quantitative studies. As detailed in the inclusion criteria, all studies reported on ED nurses' experiences with patient-related violence in their workplace; however, some studies also included details of violence from relatives or friends of patients and limited reference to horizontal violence was included in one study.27
Study sample and participants: There was a lack of homogeneity between the studies in relation to sample sizes. The sample sizes for the 11 quantitative studies ranged from 66 to 3518; with response rates from 11% to 97%. All quantitative studies recruited ED nurses; however three also included other ED employees.28,33,35Two studies undertook a nationwide survey of ED nurses.31,32The other studies recruited participants through their place of employment and the number of sites in each study ranged from one to 65.
Study design: There was homogeneity in study designs. All quantitative studies employed a cross-sectional approach and used a survey tool to collect data.
Methods/instruments: While all the quantitative studies used a survey tool to collect data; the detail provided about survey items differed between the studies resulting in a lack of homogeneity. For example, some studies did not report the length of their instrument32,33,36; others reported the number of questions in their survey, for example, Atawneh et al. used a two-part questionnaire which contained 18 questions30, and Gacki-Smith et al. used a 69-item questionnaire.31Other studies listed the number of sections on their survey which ranged from two to four. Some studies reported the types of questions used, for example, Likert-type, multiple choice and open-ended35; however the majority did not provide such information. There was also a difference between the design of questions, for example, some studies favored the use of open-ended questions, while others were multiple-choice.
Five studies reported that their survey instrument had been validated and a further two studies reported that some sections of their questionnaire had been validated. Four studies did not provide any details of validation.
Recall periods: There was a lack of homogeneity between the quantitative studies in relation to recall periods, which ranged from three and five months to up to five years. Three studies did not specifically define the recall period used, or provided very broad periods of time, for example, “…experienced throughout the career in the ED…”.34(p156). Please see Appendix V for full details.
Definition: There was a lack of homogeneity between these studies in relation to the use of a definition for “violence” and the lack of a standardized definition for the term has been acknowledged in the literature as a limitation in comparing studies. Only four of the 11 quantitative studies provided a definition of the term “violence” to their study participants; however six studies distinguished between verbal abuse and physical violence in their results.
Validation: The level of rigor and validation undertaken in the studies varied widely. Some studies provided minimal details of steps taken to ensure rigor35, while others reported significant measures taken to ensure the validity of the survey instrument used.28
Variation in study factors and outcomes measured: There was some variation between studies in the study factors and outcomes measured. A lack of homogeneity was also identified between the qualitative studies and the details are discussed below.
Context and sample: There was a lack of homogeneity between the qualitative studies in relation to sample size and the sample sizes included in this review ranged from six to 97 participants working in the ED environment.
Study sites and participants: All qualitative studies recruited ED nurses; however one study also included other ED employees.48The number of study sites varied between the qualitative studies, from one to three.
Study design: There was some variation between the qualitative studies with regards to study design. Three study designs were employed including descriptive17,45, phenomenological46, and action research.48
Methods and instruments: There were differences in data collection methods between the qualitative studies: three studies used semi-structured interviews to collect data and one conducted focus groups.
Recall periods: There was a lack of consistency in the identification of recall periods in the qualitative studies. While a recall period was defined for one study, one month17, the other three studies did not provide a definition of the recall period used.
Definition: There was also a lack of consistency between the studies in relation to the use of a definition for the term “violence”: only two studies provided a definition.
Validation: There was a lack of detail provided regarding the trustworthiness of the data analysis, in terms of dependability, confirmability, credibility and transferability across all qualitative studies. One study did not provide any detail aside from a comment about “journaling” by the author.45The other three qualitative studies provided some detail of the process used to ensure trustworthiness including member checking of thematic analysis and use of an audit trail17, and the use of three independent researchers to analyze data.48
Variation in study factors and outcomes measured: There was a lack of homogeneity between the studies with regards to study factors and outcome measures.
There was also a lack of homogeneity identified between the mixed methods studies, details of which are discussed below.
Context and sample: The sample studies varied between the mixed methods studies, with sizes ranging from 20 to 31 participants.
Study sites and participants: All mixed methods studies recruited ED nurses from one study site, with one study also including other ED employees.41
Study design and method: While all mixed method studies utilized interviews to collect data, the use of observation, digital photographs and archival records were also reported as sources of data.
Instruments: All the mixed methods studies used interview schedules and all reported that these had been piloted to establish rigor.41,53,54
Recall periods: There was a lack of homogeneity between the studies in relation to recall periods with only one study reporting on this with a recall period of six months41; the other two studies did not specify a recall period.53,54
Definition: None of the mixed methods studies included in this review provided a definition of the term “violence”; however one made the distinction between physical and verbal aggression in their results section.41
Validation: All of the mixed methods studies provided details of processes used to ensure rigor.
Variation in study factors and outcomes measured: There was a lack of homogeneity between studies in relation to study factors and outcomes measured, with some variation between studies noted.
In view of this lack of homogeneity between these studies this synthesis of the results is presented in narrative form.
Synthesis of results
Frequency of episodes of violence
Ten studies reported on the frequency of episodes of violence. Two of these were qualitative studies and the detail provided was minimal, for example, violence against nurses in the ED was a common experience, and described as a daily occurrence.46Gacki-Smith et al. reported that approximately 25% of their 3465 participants had experienced physical violence more than 20 times in the past three years, and almost 20% (n = 604) reported experiencing verbal abuse more than 200 times in the three-year study period.31Fifty nurses (70%) reported 110 episodes of violence in a five-month period: approximately five per week or two per 1000 presentations.14These episodes were perceived to be increasing in both frequency and intensity.17,48
Eight studies did not report on the frequency of episodes of violence.
In studies where nurses were only one of the professional groups studied, they emerged as the professional group at the most risk of violence.33,35For example, one study reported that nurses were the professional group most frequently targeted, experiencing 67% of the 319 episodes of physical violence in their study.35
The types of violence: physical and verbal
Fifteen studies reported on types of violence, with verbal abuse the most common type of violence experienced, reported by between 53% and 99% of those studied.14,34Swearing was identified as the most common form of verbal abuse.14,17Other types reported included yelling or using a raised voice, threatening the personal safety of staff and invading personal space41, and intimidation and harassment with sexual language/innuendo.31
Physical violence was reported by between 9% and 79% of those sampled.33,36This included overt acts such as biting, hitting, pushing, kicking and slapping14, and being spat on, hit, pushed/shoved, scratched and kicked.31Episodes of sexual harassment were also reported by some nurses, for example, 44% (n = 42) of nurses in one study reported experiencing at least one incident of sexual harassment from patients.35
The presence of weapons was reported, and this included both conventional weapons, for example, knives and guns28and opportunistic weapons, for example, a cane46, or syringes.17Two studies reported specifically that there were no episodes of physical violence with a weapon.27,30
Organizational reporting of episodes of violence
Eleven studies discussed the issue of reporting of episodes of violence with all identifying under-reporting as a common occurrence, with up to 80% of incidents of verbal and physical violence not reported.29While reporting mechanisms were acknowledged, barriers to reporting were identified including confusion about which episodes to report, time constraints and a lack of administrative support, feedback and follow-up.17,32,46Reporting systems that were not user-friendly were also cited as a reason for under-reporting17, and in one study, 85% of those surveyed were unaware of the process involved in reporting an episode of violence.27
Luck et al. reported that nurses often made judgments about specific episodes of violence which guided their decisions about whether or not to report.53The authors describe a continuum where those episodes perceived to be “within normal limits” were not reported. Mitigating patient-related factors such as anxiety, confusion and disorientation and/or a lack of appreciation or understanding of the resources and processes of the ED were described. Presentations perceived to be “appropriate” were afforded greater empathy than those deemed to be inappropriate.
Nurses who did not perceive barriers to reporting episodes of violence were much less likely to have experienced frequent ED verbal abuse than other nurses (9.78% versus 21.5% p<0.001).31Establishing a positive culture with regards to reporting of episodes of violence is an important step in creating a safe work environment.31Only 25% of participants were encouraged to report episodes of violence in one study27, and the majority of these were encouraged by colleagues, not management. Employees stated that reporting should be easy to access and quick to complete.48
Antecedents and risk factors
Eleven studies reported antecedents and precipitating factors of episodes of patient-related violence and for the purpose of this discussion these have been grouped into three main areas: patient presentation, environmental and patient behaviors as described below.
Patient-related antecedents and precipitants reported in the studies included lifestyle factors such as alcohol intoxication, substance abuse and withdrawal.14,17,45, 14,17,41Patient diagnoses and co-morbidities such as cognitive impairment, for example, dementia, delirium, head injuries and hyper or hypoglycemia. Mental health issues were also reported to be significant in a number of studies, however, this was referred to as a “mental health diagnoses” with no detail provided as to the specific condition.17,31,14, 17,31,45
Factors specific to the ED were also reported with the ED environment described as unsupportive46, and potentially contributing to the emergence of violent behavior through overcrowding, long waiting times and staff shortages.17,31The design of the ED was reported to be a contributing factor29,46, with triage identified as one the highest risk areas for potential violence.14
Some studies reported that patients often exhibited demanding behaviors and cues, and requested attention prior to exhibiting violence and these could be recognized by some staff and used to defuse potentially violent situations.14Thus less experienced staff were identified as being at greater risk of potential violence from patients.41These behaviors were described in detail by Luck et al. using the mnemonic STAMP: staring, tone, anxiety, mumbling and pacing.54
Perpetrators of violence
Fifteen studies identified the perpetrators of violence against ED nurses and patients were identified as the most common source of violence in 10 of these studies.14,17,28,31,41,45,46,48,53,54For example, in one study 319 episodes of physical violence initiated by patients were reported compared to 10 by visitors35, while patients were identified as the perpetrators in 78% of episodes (n =86) in another study.14
Family members were reported to be the most likely source of verbal abuse in pediatric EDs (82%), while episodes of physical violence were more likely to perpetrated by patients, especially those diagnosed with mental health problems.41High risk family members were identified as those with a lack of respect for women or persons in positions of authority, those under the influence of alcohol or illicit drugs and those who perceived the needs of their children were not being addressed.41
Patients' families and/or friends were more likely to be involved in episodes of violence in non-Western studies.27,29,33,34Typically this involved males and the actions were against female nursing staff. This was thought to be related to the roles played by males in these cultures and the fact that violence towards women is more prevalent in countries such as Turkey and Iran due to male dominance in these cultures.33
A distinction was made in some studies between the intent of perpetrators and the meanings behind their acts of violence.17,53Some behaviors were classified as problem behaviors by nurses rather than acts of violence. For example, episodes involving patients with dementia were not reported or viewed as seriously as other episodes of violence.35
Impact and sequelae of violence: psychological and physical effects
Twelve studies reported on the consequences of patient-related violence including physical injuries and psychological side effects which impacted on clinical practice. These were reported to be long-lasting, for example, the majority of participants in one study (n = 71, 96%) reported that they suffered from after effects lasting for up to four weeks after experiencing violence, and 93% (n =65 ) suffered from multiple effects.30Emotions reported included fear, anger and frustration41,45, and the issue of safety was addressed with nurses reported feeling unsafe and vulnerable at work.29One study reported that nurses were five times less likely to feel safe “most of the time” or “always” compared with other professional groups.28Another reported that four participants had received medical care and one psychiatric care as a result of an episode of violence and that there were three lost workdays due to injuries.35
An impact on practice was noted with nurses reporting that they avoided the perpetrators of violence and were unable to focus at work.41In another study, 37% (n=82) of the respondents had a negative total productivity score on the Healthcare Productivity Survey demonstrating decreased performance after a violent event.32One third of ED nurses in one study had considered leaving their ED or the specialty of emergency nursing as a result of violence.31
Prevention and control measures to manage episodes of violence
Twelve studies reported on prevention and control measures to manage episodes of violence and approaches perceived to be effective included education and training, provision of security and reporting. Nurses who reported that violence was an unavoidable part of their job were more likely to have experienced frequent physical violence.31Hospital policies were not perceived to support a safe ED environment46, and existing policies about conflict resolution and aggression management techniques were reported to be inadequate.48A workplace culture that tolerates violence as part of the job was also noted by Gates et al. and Gacki-Smith et al. who reported that there was an acceptance of such violence as “part of the job”.32,31
It is important to note that the effectiveness of strategies such as security, environmental controls and violence prevention training and education are difficult to extrapolate from cross-sectional data because such strategies are often initiated in EDs after violence has become an issue. This can have a confounding effect on the results and create the appearance that such strategies increase ED violence.31
Education and training
Luck et al. recommended the integration of knowledge of the STAMP behaviors into training programs to equip ED nurses with the skills to recognize and then defuse potentially violent situations.54This was reiterated by Gates et al. who discussed the need for frequent educational opportunities aimed at specific strategies to de-escalate violent situations with patients and to resolve conflicts.32The consensus was that whilst specific training was desired by nurses and perceived to be important48, it was often sporadic and fragmented in nature with a lack of consistency between trainers and programs.36A lack of training was identified as a concern in some studies45, and less than half of the EDs surveyed in one study had some type of training program for staff.28Eight-four percent of participants in another study had not received any training at all.29Similarly there was a lack of refresher programs reported36, with 64% of those surveyed by Gates et al. receiving no training in the year prior to completing the survey.35
Employees indicated that existing policies and education about conflict resolution and aggression management techniques were inadequate. They identified the need for frequent educational opportunities aimed at specific strategies to de-escalate situations with patients and resolve conflicts. Additional workplace violence education and training was perceived to be important and participants stated that this should be conducted in an interdisciplinary manner.48
The use of patient management plans to identify and flag previously violent patients was identified as preventative tool to alert staff in one study.17Another study reported the value of patient management plans, but noted that there were no systems in place to facilitate this.32
There was a perception among participants that a lack of security35, or a slow response time from security46, increased the risk of violence in the department. Employees and managers were also of the view that security and/or police presence in the ED should be increased and that guards should be armed and zero tolerance policies needed to be enforced.48Other increased security measures discussed included the use of metal detectors to screen patients and visitors for weapons28,35,46, the use of Tasers by security guards48, the use of duress alarms17, and one study discussed the introduction of a policy to restrict the number of visitors per patient.32Environmental measures such as restricted access and the use of quiet areas or safe rooms to house patients at risk of violence were also reported.32
Post episode follow-up
Six studies reported on coping strategies and follow up after episodes of violence, with all discussing deficits in this regard.17,27,29,32,34,48Management and employee participants supported the use of debriefing following episodes of violence; however they reported that this was rarely done and that support systems provided were inadequate. Barriers cited included time constraints, a lack of administrative support and a workplace culture that tolerates violence as part of the job.48Gates et al. reported that critical incident debriefing was rarely done and whilst formal debriefing was desired it was not offered in the ED.48,17,27Support was received from colleagues (40%) in the ED rather than through more formal mechanisms such as administration, and informal debriefing with colleagues was identified as the main coping mechanism in dealing with the aftermath of violence.29Participants in one study requested that a staff liaison role be created to address any ongoing concerns of the victims of violence and to ensure that their physical, emotional and psychological injuries have been addressed prior to them returning to work.48
Summary of synthesis and significant issues
Analysis of the included papers identified violence in the ED against nurses as occurring on a regular basis, in the form of verbal and physical aggression. The majority of episodes were perpetrated by patients in the department; however relatives were also identified as a risk group, particularly with regards to verbal abuse and this was more frequent in non-western ED settings.27,29,33,34The parents of pediatric patients were also identified as a group of risk.17,41Antecedents and risk factors for violence included factors related to patient presentation, for example, alcohol intoxication, substance abuse and cognitive impairment; environmental, for example, long waiting times and overcrowding; and patient behaviors, for example, STAMP behaviors. Some participants in these studies considered that a patient's presentation was a mitigating factor for their actions, for example, patients with dementia. Behaviors reported to be associated with these presentations are consistent with those described in the literature as “resistance to care” behaviors55, and represent defensive responses rather than offensive aggression and violence.56
Exposure to violence led to physical and psychological consequences for nurses and these were often long-lasting in nature and affected the quality of care provided to patients.17,32,41Despite the frequency and severity of episodes of violence, reporting of these events was not the norm and their subsequent management was perceived to be largely inadequate.31,35Risk management centered on training in de-escalation techniques and security measures such as restricted access to the department and the presence of security guards.30,31,35,36Violence against ED nurses was identified as an area of increasing concern and one in need of more action in terms of management of the issue and further research into the phenomenon.
Limitations of the review
All of the studies considered for this review contained methodological shortcomings. Eleven of the included studies used a cross-sectional survey to collect data and of these the majority used non-validated instruments.27,28,31,33-35Four of these studies reported sample sizes of less than 10014,30,34,36, and some studies reported response rates of as low as 9%, which may compromise the external validity of the study, or the degree to which the findings can be generalized to other populations or environments.32
All studies collected self-report data, which is subject to recall bias. Such data is subjective in nature and influenced by the participants' perceptions of events and the time period involved. There is the potential for inaccuracy in the data collected; however, self-report data is necessary in this area of study due to the under -reporting of episodes of violence.57While some studies attempted to minimize this by limiting the period of recall to between one and 12 months, others included no such limitations. The difference in data collection periods between the studies was a limitation making cross comparison difficult. In addition, the anonymous nature of the surveys in the quantitative studies meant that there was no way to identify whether participants were similar to non-participants in relation to their experiences with violence. Gordis reports that generally people who do not participate in a research project differ from those who do in regard to many characteristics.58This means that the results of these studies may have been influenced by self-selection bias. The use of convenience sampling and the use of only one study site in some studies may weaken the transferability of findings in qualitative studies and generalizability of results in quantitative studies. There is also the potential for bias because those interested in the topic or who had experienced patient-related violence may be more likely to participate in studies on the topic which could result in a biased view of the topic.58
The lack of a consistent definition of the term “violence” by each study makes direct comparison between the studies more difficult. Only five of the 18 included studies defined “violence”, while a further six studies distinguished between verbal abuse and physical violence in their results; however, they did not provide a definition of either term. Many of the quantitative studies employed basic descriptive statistics and results were only reported in raw numbers and percentages and did not report statistically significant results.
Aspects of trustworthiness and rigor were addressed by some studies; however, other studies did not provide any details or provided minimal information. Member checking was conducted by two of the four qualitative studies; however, in the case of Early and Hubbert it was unclear how many participants were involved in this process.46Both mixed methods studies by Luck et al. also employed member checking which demonstrates credibility and dependability.53,54In addition, they pilot tested the interview schedule. Audit trails, which help to demonstrate confirmability through the documentation of the research process followed, and enable future researchers to replicate the study, were used in one qualitative and one mixed method study.
Methods employed in the quantitative studies to ensure rigor ranged from the use of expert panels and testing groups to statistical testing using the test-re-test method to determine reliability and Cronbach's Alpha for measurement of item constructs. Four studies employed statistical methods on all or some aspects of their survey tools and three employed experts to either pilot test or review their survey tool. Four studies did not provide any details of the survey development process. The lack of consistency in the methods employed made comparison of the results more difficult.
Violence from patients and/or their friends and family has emerged as a significant threat to the safety of ED nurses in their daily working lives. Nurses are exposed to levels of physical and verbal violence that are so high they have become an expected part of the job for many and are tolerated on a daily basis. However, such episodes of patient-related violence can have long-term physical and psychological consequences for nurses and have resulted in the death of a number of nurses on the job. There is also a flow-on effect on their professional lives which can impact negatively on the level of care afforded to future patients and potentially compromise patient safety. Nurses feel unsupported by management in terms of support and follow-up offered to them following episodes of violence. This has led to chronic under-reporting of episodes and a discrepancy between the actual incidence of violence experienced by nurses in the ED and official figures. Healthcare employers are obligated to provide a safe working environment for all who enter their facilities under the Work Health and Safety Legislation; however, the current reality is that nurses frequently feel unsafe at work and this has the potential to impact negatively on patient safety. Whilst it is not possible to eliminate the threat of violence from high risk environments like the ED, employers must act to provide a safer environment where this threat is minimized and they should support their staff in the event of an episode occurring.
Implications for practice
Research should continue into violence against ED nurses to provide the evidence needed to inform management and government and guide policy development. Multi-site studies with large sample sizes could be conducted to add support to the body of the evidence on the topic. Existing studies all have limitations due to recall bias; however, the use of archival records to measure episodes of violence is likely to fail given the chronic under-reporting identified in this review. There is a need for prospective studies to be conducted to collect data on episodes of patient-related violence as they occur in real-time.
Implications for research
Research should continue into violence against ED nurses to provide the evidence needed to facilitate better prevention and management of this issue.
There is minimal evidence about the extent and frequency of episodes and resultant injury of patient-related violence in the ED and of the experiences and perceptions of ED nurses in Australia.14,17,53,54,59,60Urgent preventative measures and responses to patient-related violence are required to address the phenomenon to ensure that ED nurses are working in a safe environment.
Conflict of Interest
The authors declare no conflict of interest.