Secondary Logo

Journal Logo


Quality Improvement Project to Manage Workplace Violence in Hospitals

Lessons Learned

Schwartz, Franklin DNP, APRN, PMHNP-BC; Bjorklund, Pamela PhD, APRN, PMHNP-BC, CNS-BC

Author Information
Journal of Nursing Care Quality: April/June 2019 - Volume 34 - Issue 2 - p 114-120
doi: 10.1097/NCQ.0000000000000358


Workplace violence in hospitals is pervasive, underreported, and tolerated and has deleterious effects on the physical and emotional well-being of health care staff.1–3 Violence that is perpetrated by customers of an establishment (ie, type II workplace violence) is common in hospitals, with an overall prevalence ranging from 31% to 53% in hospital-based studies.2 Little is known about the scope of the problem in office-based practices.3 Between 2011 and 2013, the number of workplace assaults averaged approximately 24 000 annually, with nearly 75% occurring in health care settings.3 Relative to all other occupational groups, health care professionals are at a greater risk for experiencing work-related violence.2–4 Approximately 50% of health care professionals in general medical settings have experienced verbal violence from patients or visitors, and up to 25% have been subjected to physical violence.5 Rates up to 100% for verbal assault and 82% for physical assault have been reported in high-risk settings (eg, emergency department [ED] and inpatient psychiatric units).3 Up to 95% of nurses have experienced either verbal or physical violence at some point during their careers.4

Since rates of assault correlate with patient contact time, nursing staff and provider groups are victimized most often.2,3 They are 7 times more likely to be exposed to physical violence and twice as likely to be exposed to nonphysical violence compared with all other at-risk health care workers.1 While most occupational groups involved in direct patient care are at a notable risk of violence, nurses have the highest prevalence, followed by nurses' aides and providers. Overall, 82% to 100% of ED nurses, 70% to 99% of psychiatric staff, 61% of home health care workers, 59% of nursing home aides, 33% of pediatric residents, and 25% of ED physicians have suffered type II violence.3 Workplace violence has a psychological impact on health care staff, and the costs of stress and violence together account for approximately 0.5% to 3.5% of the gross domestic product (GDP) per year.6 Interventions to manage violence risk are needed urgently. The purpose of this article is to describe one such effort.


The study of workplace violence has focused to date on quantifying the problem and profiling perpetrators and victims. Intervention studies are scarce, and no universally applicable training program currently exists to reduce patient and visitor violence against health care staff.3 Unlike psychiatric settings, lack of evidence-based understanding of violence has persisted in general medical settings.4 Moreover, the aggression of patients or visitors toward health care staff is complex, varies from verbal aggression to physical assault, and stems from multiple factors, including the (a) prevalence of lethal weapons owned and carried by patients, families, and/or friends; (b) increased use of hospitals by police and the criminal justice system for the care of acutely disturbed, violent individuals; and (c) increasing numbers of acute and chronically mentally ill patients who may be released from hospitals without adequate follow-up.7 Other factors include long waits, inadequate staffing, and lack of staff training in the recognition and management of escalating aggression.7

Patients tend to be more violent than visitors.3,4 Patients most likely to exhibit aggressive behaviors in general medical areas often have histories of violent behavior, psychotic disorders, cognitive impairment, confusion or delirium, alcohol or drug withdrawal, and mental illness.4 Patients who suffer from mental illness often have other characteristics that put them at even greater risk for violence, including history of trauma, unemployment, homelessness, and experience with the criminal justice system.8 Unlike psychiatric nurses, who are trained in de-escalation techniques to manage aggression, nurses employed in general medical areas are not always prepared to manage it.9


The Bureau of Labor and Statistics found that 48% of all nonfatal injuries to nurses from occupational assaults happened in a hospital setting.2,10 Annual missed workdays related to patient assaults are costly, and workers compensation claims are soaring.11 In 2011, the number of days away from work due to staff injuries and illnesses in health care venues was estimated at 14.6 days per 10 000 workers compared with 3.8 days per 10 000 workers in the private sector.8 Although the majority of physical injuries tend to be minor, career-ending incidents involving permanent disability are not uncommon, and psychological wounds can persist for months or years.12 Up to 61% of assaulted staff report symptoms of posttraumatic stress disorder and commonly feel a loss of confidence in their professional competence and appropriateness for the job.11

Violence creates a hostile work environment and negatively affects organizational outcomes in various ways, including decreased morale, increased job stress, increased staff turnover, and decreased trust in management and coworkers.13 Prevention and management strategies to reduce patient and visitor violence in general medical areas must include appropriate training.14 The Occupational Safety and Health Administration (OSHA) recommends that training programs include topics such as risk factors for violence, early recognition of behaviors that may escalate to assaults, means to prevent or diffuse volatile situations, and information on multicultural diversity to increase staff sensitivity to racial and ethnic differences.7 Training should be patient-centered with an emphasis on building relationships by engaging in empathic interactions with patients and visitors.14 The diverse nature of the violence highlights the need for broad training in risk recognition and de-escalation.2


The purpose of this project was to pilot a violence management training program consistent with OSHA recommendations in a general medical unit, which was selected on the basis of both the increased cognitive impairment of patients often admitted to this unit and the demand from the unit's staff for measures to protect them from patient and visitor violence.


The methodology included an independent pre/posttest design to measure improvement in participant knowledge following an educational program that included in situ simulation training in de-escalation techniques. The Staff Observation Assessment Scale Revised (SOAS-R) was selected for data collection on aggressive behaviors pre- and postimplementation of the violence management program.15–21 An institutional review board reviewed the project proposal and declared it exempt.

Setting and sample

The violence management training program was implemented at a large teaching hospital in a Midwestern state on a 39-bed general medical unit that provided care to patients on stroke/cerebrovascular, epilepsy, general neurology, and family medicine services. The patient population served by this unit ranged in age from 18 years to more than 90 years. Health care staff who worked on this unit had not received any other training related to violence prevention during their employment orientation. Physicians and other provider groups were excluded, because they were not responsible for patient and visitor milieu management. Ultimately, only 79.4% (n = 93) of the unit's health care staff attended the training and were counted as participants, including 65 (70%) associate or bachelor's prepared registered nurses, 2 (2%) masters prepared advanced practice registered nurses, and 26 (28%) patient care attendants.


In collaboration with the unit's Nursing Education Specialist, the project leader developed the violence management educational material to align with OSHA recommendations and best evidence on early recognition of signs of escalation, early identification of risk factors for violence, and relationship building with the recipients of care.7,13,22–26 Participants received this content along with orientation to the SOAS-R in three 2-hour education sessions. Following orientation to the SOAS-R, participants were tasked to collect baseline and postintervention data with the SOAS-R for 3 months before and 3 months after receiving violence management training. Participants were to complete the SOAS-R after witnessing or otherwise encountering aggressive patient or visitor behavior, defined as any verbal, nonverbal, or physical behavior that either threatened or actualized harm to self, others, or property.

The first half of the training included review of the phases of the assault cycle and identified potential escalating behaviors during each phase along with appropriate verbal and nonverbal communication strategies for de-escalation of phase-related behaviors. Early recognition of arousal onset was emphasized along with limit-setting and person-centered care to meet patient needs proactively. Examples of aggressive scenarios were provided, and a short video clip illustrated how health care staff members escalate or de-escalate a situation, depending on their own verbal and nonverbal communication. The second half of the training program applied the content to simulated scenarios of actual unit incidents, using 3 actors hired from the organization's simulation center. Participants volunteered to role-play each scenario, while other participants observed and provided feedback. Six 4-hour training sessions were required to accommodate all participants.

Measures and analysis

The SOAS-R is an incident-based instrument to record discrete episodes of aggressive behavior. It has good psychometric properties, is recommended in clinical studies of inpatient aggressive behavior, and can illuminate the specific factors connected to violent outbursts.18–21 It records data related to provocation, warning signs of increased risk of aggression, means of violence used by patients, targets of aggression, consequences to victims, and measures to stop aggression. It includes a common visual analog scale to measure participants' perceptions of the severity of aggressive behaviors, ranging from 0 (not severe at all) to 10 (extremely severe). The violence management program also included a 5-question pretest/posttest to evaluate participants' knowledge of program content. Both tests were identical and were completed anonymously. Questions pertained to 2 scenarios role-played during the training. The project leader administered and collected pretests at the beginning of each session. The posttest was administered electronically to each participant 3 months after the training session. Data analysis included a Z-score 1-tailed test to look for significant differences between pre- and posttest scores.


Of 117 unit staff members, 79.4% (n = 93) attended the violence management training program. Among attendees, 100% completed the pretest and 86% (n = 80) completed the posttest, which was less than the targeted outcome of 100%. Z-score analysis showed that posttest scores were not significantly different from pretest scores for 2 of the questions. However, posttest scores were significantly higher than the pretest for questions 3 (precipitating factors to acting-out behavior) (P < .0003), 4 (purpose of a supportive stance) (P < .0001), and 5 (keys to setting limits) (P < .0001) (Table).

Table. - Knowledge About Violence Management
Questions Pretest, n (%) Posttest, n (%) Difference, % P
1. What level of behavior is the patient exhibiting? 58 (62.37) 55 (68.75) 6.38 .1894
2. What would be an appropriate therapeutic response to help decelerate the escalating behavior? 64 (68.82) 63 (78.75) 9.93 .0708
3. List 2 precipitating factors that might contribute to acting-out behavior. 73 (78.49) 77 (96.25) 17.76 .0003
4. What is the purpose of a supportive stance? 12 (12.90) 67 (83.75) 70.85 <.0001
5. List 3 keys to setting limits. 9 (8.60) 49 (61.25) 52.65 <.0001

For reasons explored in the “Discussion” section of this article, only 1 SOAS-R was completed during the first 3 months and 5 were completed during the 3-month posttraining period. A single SOAS-R report during the preintervention phase of the project was insufficient to constitute a baseline measure. All collected forms were completed correctly. An analysis of visual analog scores from the 6 completed SOAS-R forms revealed a mean incident severity score of 7.16, indicating that documented aggression was perceived as severe. Insufficient data posed a serious limitation to the project and failed to show a decrease in the frequency, nature, and severity of violence. Nevertheless, the SOAS-R provided valuable information about each of the documented incidents of aggression (see Supplemental Digital Content, Table, available at:


This quality improvement (QI) project met its overall goal of piloting a violence management training program for the health care staff on a general medical unit. Although the project at its inception reached only 79.4% (n = 93) of the unit staff, participants demonstrated on the posttest significantly increased knowledge of measures required to manage escalating, aggressive, and actual or potentially violent behavior. Consistent with background literature, violent incidents presumably were underreported, and no conclusions can be drawn about the effect of the program on the rate of violence in general medical settings. While 100% of documented SOARS-R forms were completed correctly, data were insufficient to demonstrate that the violence management program reduced the number of aggressive incidents that occurred on the unit during the postintervention period, nor did the project demonstrate a clear increase in participants' confidence in managing patient or visitor aggression, as this was not directly measured. However, important lessons were learned for the ongoing process of continuous QI throughout all phases of this project.

Lessons learned


Not all of the expected health care staff members attended one of the scheduled violence management training sessions. Multiple factors may have contributed to less than 100% attendance, including staff vacancies, illness, scheduled vacations, patient acuity, and short staffing. Moreover, a 100% target for attendance might be unrealistic in the absence of organizational and structural support for mandatory training of all at-risk staff members. An expected outcome of 80% is feasible and more realistic. This project achieved a 79.4% attendance outcome for its violence management pilot program. Given the vulnerability of the provider occupational group, efforts should be made to enlist providers in future trainings. Ideally, administrative support would make it possible for all at-risk staff members to attend future mandatory trainings at regular intervals.

Pre- and posttests

While 100% of participants completed the pretest, which was offered on-site during the educational program, 86% of participants completed the posttest, which the facility's online education program sent electronically to each participant 3 months after the training. Resignations or terminations may have factored into the return rate. Although the posttest was included as part of participants' online education plans, participants were not required to complete the posttest within a certain time frame, and reminders were not sent. While a posttest administered online 3 months after receipt of training may demonstrate the durability of learning, it may be too long a time period to expect a 100% return rate. An expected outcome of 80% is more realistic for a posttest administered that long after initial training. This project's 86% return rate 3 months posttraining is good and can be attributed to its integration into participants' individual online education plans.

Notably, results showed important differences from pre- to posttest in some of the 5 test questions. Answers to questions about the level of aggressive behavior and appropriate therapeutic response were not significantly different from pre- to posttest. Health care staff may have been experienced already in recognizing triggering events and behavioral changes and in providing the appropriate responses for that level of behavior. However, significant differences between the other 3 questions were observed from pre- to posttest with respect to recognition of contributing factors that might precipitate aggressive behavior, how a supportive stance promotes safety and respect, and effective ways to set limits.

Frequency, nature, and severity of violence

Consistent with results of a recent review of nursing literature, training showed no clear association with reductions in the incidence of violence perpetrated by patients.27 Data collected from the SOAS-R during pre- and postintervention periods were insufficient to draw any conclusions about associations between violence management training and the rate of aggressive incidents on the medical unit. Despite ongoing monitoring and encouragement by the project leader, the staff completed only 1 form during the pretraining period and 5 forms during the posttraining period. It is possible the health care staff saw acts of aggression only in terms of physical threat and did not document less severe acts of aggression, which is a common form of underreporting and is consistent with extant literature that shows pervasive patterns of underreporting of violent incidents across health care occupational groups and across settings.2,3,28 This is unfortunate, given that nonphysical forms of violence constitute a risk factor for physical assault.23 Given possible issues of short-staffing, patient acuity, heavy documentation requirements, other unknown workflow-related issues, and the inherent stress of confronting patient aggression, another hypothesis is that the SOAS-R was simply not a priority compared with other aspects of patient care. The extra steps of accessing the form online and returning it in hardcopy may have increased the perceived workload.

Although the data were insufficient to show any associations between violence management training and decreased frequency of violent incidents, the data that were gleaned from the SOAS-R provided beneficial information for organizational leadership (see Supplemental Digital Content, Table, available at: Health care staff members were the target of aggression on all completed SOAS-R forms, and they felt threatened on 100% of the reported incidents of aggression. Aggressive incidents on the medical unit escalated to physical violence and caused physical injury to health care staff in 33% of reported incidents. Health care staff engaged with the aggressive patient and attempted to talk to the patient during 100% of the reported incidents. When behaviors reach the level of escalation, verbal interventions must change in nature from supportive to limit setting. If verbal interventions fail to provide the required structural limits, verbal and physical aggression can escalate to violence and injury.


Other limitations to this project included that not all health care occupational groups at risk for type II violence were participants in this QI project, the content validity of the pre/posttest was not established, test scores for individual participants were not paired, and a 5-question test may not have been sufficiently robust to measure comprehensive learning. By far, however, the most severe limitation to this project was constituted by the barriers that prevented SOAS-R completion, which was not a mandatory process. In addition, health care staff members throughout the organization frequently used hospital security staff and the Behavioral Emergency Response Team (BERT) team to manage escalating patient and visitor behavior. Although they have expertise in violence prevention and management of aggression, security staff and BERT personnel were not project participants, received no SOAS-R training, and did not complete SOAS-R forms.


Health care organizations are not required by law in most states to have specific violence prevention strategies in place, and The Joint Commission has only vague policy requirements that are open to interpretation.3 While no universally applicable training program currently exists to reduce violence in health care settings, training is still a necessary, albeit insufficient, measure to manage and mitigate the ongoing threat of violence. The OSHA has provided guidelines to reduce violence risk in health care settings; however, the onus to voluntarily implement customized training programs falls upon the leadership of health care organizations.

This QI project demonstrated that violence management programs are feasible on general medical units and can enhance the knowledge of unit staff to manage aggression and violence more effectively. Insufficient data precluded conclusions about the program's capacity to reduce the incidence of aggression on such units. However, the SOAS-R is a valuable tool for collecting data on the frequency, nature, and severity of aggressive incidents; and barriers to its use should be mitigated. Adjustments to the data collection processes to improve rates of SOAS-R completion could include better project management with additional completion reminders, clarifying lines of responsibility for documentation, simplifying access to and submission of the form, linking SOAS-R completion to existing, mandatory incident reporting, and involving security and existing BERT personnel in aggressive incident training, tracking, and reporting. While in situ training may enhance staff confidence to manage aggressive incidents, intervention studies are needed to more effectively translate results into training programs and systems improvements.


1. Magnavita N. Workplace violence and occupational stress in healthcare workers: a chicken-and-egg situation—results of a 6 year follow up study. J Nurs Scholarsh. 2014;46(5):366–376.
2. Pompeii LA, Schoenfisch AL, Lipscomb HJ, et al. Physical assault, physical threat, and verbal abuse perpetrated against hospital workers by patients and visitors in six U.S. hospitals. Am J Ind Med. 2015;58(11):1194–1204.
3. Phillips JP. Workplace violence against health care workers in the United States. N Engl J Med. 2016;374(17):1661–1669.
4. Hahn S, Muller M, Needham I, et al. Factors associated with patient and visitor violence experienced by nurses in general hospitals in Switzerland: a cross sectional survey. J Clin Nurs. 2010;19(23/24):3535–3546.
5. Hahn S, Zeller A, Needham I, et al. Patient and visitor violence in general hospitals: a systematic review of the literature. Aggress Violent Behav. 2008;13(6):431–441.
6. Kim S, Idecker K, Todicheeney-Mannes D. Usefulness of aggressive behavior risk assessment tool for prospectively identifying violent patients in medical and surgical units. J Adv Nurs. 2012;68(2):349–357.
7. Occupational Safety and Health Administration. Guidelines for Preventing Workplace Violence for Health Care and Social Service Workers. Washington, DC: Government Printing Office; 2016. OSHA 3148-06R 2016.
8. Lipscomb JA, Ghaziri ME. Workplace violence prevention: Improving front-line health-care worker and patient safety. New Solut. 2013;23(2):297–313.
9. Kynoch K, Jung Wu C, Chang AM. Interventions for preventing and managing aggressive patients admitted to an acute hospital setting: a systematic review. Worldviews Evid Based Nurs. 2011;8(2):76–86.
10. Hader R. Workplace Violence Survey 2008. Nurs Manag. 2008;39(7):13–19.
11. Hankin CS, Bronstone A, Koran LM. Agitation in the inpatient psychiatric setting: a review of clinical presentation, burden and treatment. J Psychiatr Pract. 2011;17(3):170–185.
12. Needham I, Abderhalden C, Halfens R, et al. Non-somatic effects of patient aggression on nurses: a systematic review. J Adv Nurs. 2005;49(3):283–296.
13. National Institute for Occupational Safety and Health. Violence: occupational hazards in hospitals. NIOSH Publication No. 2002-101. Published 2002. Accessed March 21, 2018.
14. Hahn S, Hantikainen V, Needham I, et al. Patient and visitor violence in the general hospital, occurrence, staff interventions and consequences: a cross sectional survey. J Adv Nurs. 2012;68(12):2685–2699.
15. Palmstierna T, Wistedt B. Staff observation aggression scale, SOAS: presentation and evaluation. Acta Psychiatr Scand. 1987;76(6):657–663.
16. Nijman HL, Muris P, Merckelbach HL, et al. The Staff Observation Aggression Scale Revised (SOAS-R). Aggress Behav. 1999;25(3):197–210.
17. Nijman HL, Palmstierna T, Almvik R, et al. Fifteen years of research with the Staff Observation Aggression Scale: a review. Acta Psychiatr Scand. 2005;111(1):12–21.
18. Tenneij NH, Goedhard LE, Stolker JJ, et al. The correspondence between the Staff Observation Aggression Scale-R and two other indicators for aggressive incidents. Arch Psychiatr Nurs. 2009;23(4):283–288.
19. Kobes M, Nijman H, Bulten E. Assessing aggressive behavior in forensic psychiatric patients: validity and clinical utility of combining two instruments. Arch Psychiatr Nurs. 2012;26(6):487–494.
20. Iennaco JD, Whittemore R, Dixon J. Aggressive event incidence using the Staff Observation of Aggression Scale-Revised (SOAS-R): a longitudinal study. Psychiatr Q. 2017;88(3):485–499.
21. Paschali M, Kamp D, Reichmann C, et al. A systematic evaluation of impulsive-aggressive behavior in psychogeriatric inpatients using the Staff Observation Aggression Scale-Revision (SOAS-R). Int Psychogeriatr. 2018;20(1):61–68.
22. Kaplan SG, Wheeler EG. Survival skills for working with potentially violent clients. Soc Casework. 1983;64(6):339–346.
23. Lanza ML, Zeiss RA, Rierdan J. Non-physical violence: a risk factor for physical violence in health care settings. AAOHN J. 2006;54(9):397–402.
24. Ford K, Byrt R, Dooher J. Preventing and Reducing Aggression and Violence in Health and Social Care. Keswick, England: M & K Publishing; 2010.
25. Richmond JS, Berlin JS, Fishkind AB, et al. Verbal de-escalation of the agitated patient: consensus statement of the American Association for Emergency Psychiatry project BETA de-escalation workgroup. West J Emerg Med. 2012;13(1):17–25.
26. Butera V. Cal/OSHA proposes extensive regulations to address workplace violence in health care settings: will federal OSHA follow suit? Employee Relat Law J. 2016;42(2):47–49.
27. Heckemann B, Zeller A, Hahn S, et al. The effect of aggression management training programmes for nursing staff and students working in an acute hospital setting: a narrative review of current literature. Nurse Educ Today. 2015;35(1):212–219.
28. Pompeii L, Dement J, Schoenfisch A, et al. Perpetrator, worker and workplace characteristics associated with patient and visitor perpetrated violence (Type II) on hospital workers: a review of the literature and existing occupational injury data. J Safety Res. 2013;44:57–64.

aggression; hospitals; risk management; violence management; workplace violence

Supplemental Digital Content

© 2018 Wolters Kluwer Health, Inc. All rights reserved.