Secondary Logo

Journal Logo

CASES OF NOTE

Nonviolent Crisis Intervention Training and the Incidence of Violent Events in a Large Hospital Emergency Department

An Observational Quality Improvement Study

Gillam, Sally Wakefield DNP, BSN, MAHS, RN, NEA-BC

Author Information
Advanced Emergency Nursing Journal: April/June 2014 - Volume 36 - Issue 2 - p 177-188
doi: 10.1097/TME.0000000000000019
  • Free

Abstract

VIOLENT EVENTS in health care workplaces pose threats to patients and caregivers and complicate the delivery of patient care. The National Institute for Occupational Safety and Health (NIOSH) defines workplace violence as “violent acts (including physical assaults and threats of assaults) directed toward persons at work or on duty” (2002, p. 1).

Nursing staff who work in emergency departments (EDs), mental care health settings, intensive care units, obstetrical/gynecological, and medical/surgical areas have a greater risk of experiencing workplace violence than do nurses in other areas (Gillespie, Gates, & Berry, 2013). Anxiety may partly induce violent behaviors by patients, family members, or incidental visitors, which would not otherwise be demonstrated in EDs (Luck, Jackson, & Usher, 2008). Underlying causes may also be attributable to stress induced by illness or injury (Gates, Ross, & McQueen, 2006). Additional factors may influence violent events, including alcohol use, drug use, and organizational characteristics (Gacki-Smith et al., 2009; Gillespie et al., 2013). Improved understanding of these factors may help caregivers positively influence potentially violent individuals, diffuse potentially violent situations, and increase the quality and safety of emergency care for patients (Gates, Gillespie, & Succop, 2011).

LOCAL PROBLEM

At the author's facility, patient care is interrupted when ED staff members are challenged by patients, family members, or other ED visitors who become frustrated or combative. In situations where violence appears imminent or when violence has already occurred and an assailant needs to be subdued, caregivers declare code purples via the public address system. Code purples initiate emergency responses by the hospital security team.

In mid-2011, concerns about employee morale and personal safety spurred investment in a nonviolent crisis intervention (NCI) training program for hospital employees. All hospital employees were eligible to attend the training. The 8-hr NCI training curriculum included the following items (Crisis Prevention Institute, 2013):

  • How to identify crisis-related behaviors.
  • How to respond to each behavior to prevent escalation.
  • Use of verbal and nonverbal techniques to defuse hostile behavior and resolve a crisis before it becomes violent.
  • How to cope with one's own fear and anxiety.
  • How to avoid injury if behavior becomes physical.

A quantitative quality improvement (QI) study was initiated in late 2012 to evaluate the effectiveness of the ongoing NCI training investment. The QI study focused on the ED, because during the 6 months prior to the study, 57.8% (n = 64) of code purples at the hospital occurred there.

INTENDED IMPROVEMENT

The study intended to evaluate the NCI training, its impact on reducing violent events in the ED, and the training investment. Any significant change in code purple incidence would be examined in terms of cost versus benefit and used to evaluate the investment decision.

STUDY QUESTION

What benefits are derived from providing NCI training to ED personnel to reduce violent events that are manifested as code purples?

METHODS

The approach used a single-phase observational study for a 1-year period. The study began at 12:01 a.m. central standard time November 1, 2012, and ended at 11:59 p.m. central standard time October 31, 2013. Data were reviewed in real-time to allow proactive adjustment of security staffing to ensure adequate security coverage during anticipated high traffic periods.

Ethical Issues

Participants for this study were anonymous by design. Formal signed approval for obtaining the data and completing the study was obtained from all members of the hospital quality council and the hospital chief executive officer.

Setting

The primary ED of a 304-bed acute tertiary care hospital served as the study setting. The primary ED treats more than 75,000 patients a year. The ED also functions as a staging area for a considerable number of psychiatric patients (average number greater than eight per day), many of whom must be sustained for indefinite periods while psychiatric facilities struggle to find available beds. All ED employee roles were included in the NCI training: registered nurses, patient care technicians, unit clerks, paramedics, administrative assistants, and managers. Physicians (nonemployees of the hospital) were not included in the training because of corporate restrictions at the facility regarding ethical guidelines and business practices.

Planning the Intervention

At the start of the QI study, more than 40% of the ED staff members were already trained in NCI. The study was to determine the effects of ongoing NCI training on the frequency of code purples in the ED. Any significant change in the number of code purples would be examined in terms of NCI training completed during the study period. The findings, if favorable, would be used to justify ongoing NCI training to minimize future code purples.

The study implemented repeatable processes at the department level for capturing and reporting (a) code purples and (b) training activity already in progress. The reporting structure avoided increased record keeping by caregivers. This approach also avoided skewing the data due to violence underreporting, which has been widely documented in violence studies (Ben Naten, Hanukayev, & Fares, 2011; Gacki-Smith et al., 2009; Gates et al., 2011; Medley et al., 2012; Sato, Wakabayashi, Kiyoshi-Teo, & Fukahori, 2013). This data collection approach also avoided the Hawthorne effect (i.e., changes in staff behavior induced by their awareness of the data being collected; Leonard, 2008; Mayo, 1933). The hospital security team provided weekly reports on code purple activity. Human resources–related and training-related activity from the ED was reported monthly. Additional data regarding volume indicators were provided by the accounting department.

Planning the Study of the Intervention

If skills for identifying and de-escalating crisis-related behaviors were applied successfully by trainees, NCI training was expected to result in fewer code purples per ED visit. Anticipated confounding factors were identified to be addressed if confounding was confirmed. The expected ED confounding factors were the psychiatric patient mix (Gillespie et al., 2013), staff turnover/experience (Cahill, 2008; Sato et al., 2013), staff gender mix (Ben Naten et al., 2011), and ED wait times (Gates et al., 2011; Gillespie et al., 2013). A subsequent regression plot was planned to determine the incremental effect of the independent variable on the dependent variable. The variables analyzed were as follows:

  • The percentage of cumulative NCI-trained ED staff by month.
  • The percentage of ED staff completing NCI training each month.
  • Code purple events per ED visit per month.

The expected confounding variables were as follows:

  • The percentage of ED psychiatric patients by month.
  • The percentage of male versus female ED staff.
  • Average ED wait-to-greet times each month.
  • The percentage of ED staff who left employment in the current month, the previous month, or the current 60-, 90-, 120-, 150-, and 180-day periods.

If no significant negative correlations were found between cumulative NCI training and code purple incidence, additional training variables were to be examined:

  • The percentage of ED staff completing NCI training in the current month, the previous month, and the current 60-, 90-, 120-, 150-, and 180-day periods.

Methods of Evaluation

Data were consistently collected and analyzed to ensure their usefulness to the study. Periodic validation occurred; occasional inquiries were made when code purple activity spiked or when activity seemed to decrease. Spikes were almost always valid. For most exceptions, code purples had been cancelled but the cancellations were not captured in the security log. In those instances, reconciliation occurred between the ED leadership and security staff on the same day and reports were updated.

Analysis

IBM SPSS Statistics Version 22 (IBM Corporation, 2013) was used to examine the data. Analysis used a simple two-variable model, with one independent variable and one dependent variable.

  • Independent variable: Cumulative percentage by month of ED staff completing NCI training
  • Dependent variable: Monthly frequency of ED code purples

Correlation between the independent and dependent variables was calculated using Pearson's r. A line chart was used to depict the monthly code purple events per 1,000 ED visits in comparison with the percentage of NCI-trained ED personnel. A linear regression plot determined the incremental impact of NCI training on the code purple incidence. The numerical results were compiled and used to evaluate the cost-effectiveness of training.

RESULTS

During the study period, 111 ED code purple incidents occurred across 76,246 ED visits, a mean rate of 1.46 events per 1,000 patient visits. Emergency department code purple events accounted for 55.2% (n = 111) of all code purples at the facility during the period. The ED reported nearly triple the number of the next highest reporting unit (telemetry unit).

ED Code Purple Frequency by Date

Monthly code purple incidence varied between 0.56 and 2.68 events per 1,000 ED visits. Monthly activity is shown in Table 1.

Table 1
Table 1:
Monthly code purple activity

NCI Training for ED Staff

At the start of the study, 42% of ED staff members (n = 104) had completed NCI training. Upon completion of the study, 75% of ED staff members (n = 109) had completed NCI training. Of those who left employment, 37% (n = 24) completed NCI training before leaving. The training progress through the study period was not uniform due to fluctuations in staff hiring, attrition, and scheduled training. The cumulative progression of ED NCI training is shown by the solid line in Figure 1.

Figure 1
Figure 1:
Code purples per 1,000 visits versus the proportion of staff trained.

Code Purple Incidence Versus NCI Training

Figure 1 shows a line chart comparing the ED code purple incidence per 1,000 patient visits by month with the cumulative proportion of NCI-trained ED staff. Both curves demonstrate trend lines with positive slopes. This behavior was not expected; code purples were expected to decrease in response to progressively greater percentages of NCI-trained ED staff. Further study was needed, so analysis was performed on the additional variables, that is, the percentages of ED staff trained in the current 60-, 90-, 120-, and 150-day periods.

The additional analysis yielded a strong, negative, highly significant correlation between monthly code purple incidence and the percentage of ED staff trained in NCI within the current 90-day period (r = −0.756, p = 0.004). The regression plot for the relationship is shown in Figure 2.

Figure 2
Figure 2:
Regression plot. Code purples versus 90 days staff trained.

Another strong, highly significant correlation was found between monthly code purple incidence and the monthly percentage of psychiatric patients in the ED (r = 0.824, p = 0.001). A moderate, significant correlation was also noted between monthly code purple incidence and the percentage of cumulative staff turnover in the current 90-day period (r = 0.594, p = 0.042). The correlation matrix showing the variables of interest is given in Table 2. The additional analysis identified significant negative correlations between monthly code purple incidence and both the 120-day and 150-day percentages of NCI-trained ED staff. Those correlations were nearly as strong as the 90-day figure. No correlations between the cumulative percentage of NCI-trained staff and the number of code purples were found for periods greater than 150 days.

Table 2
Table 2:
Pearson's r correlation matrix—Variables of interest

Adjustments addressed confounding effects from the psychiatric patients and 90-day employee turnover. A partial Pearson's r used the psychiatric and 90-day turnover rates as control variables. The partial correlation continued to reveal a strong, negative Pearson's value (r = −0.675, p = 0.032) between code purples and the percentage of ED staff trained in NCI within the past 90 days. The partial linear regression formula provided by SPSS is as follows:

Changes in Processes and Outcomes Associated With the Intervention

The proportion of ED staff increased by 4.8% (n = 104) during the study period. Security staffing adapted for temporary intervals during the study period. The security adaptations ensured that sufficient security staff members were available for anticipated high traffic periods (weekends that coincided with end of the month, various local events). Overall, however, the level of security staffing did not increase.

DISCUSSION

The intent of the training was to defuse a greater number of potentially violent situations, thus lowering the code purple incidence. When greater percentages of staff were trained in NCI in the previous 90–150 days, monthly code purple incidence decreased. This contrasted with code purple incidence over the yearlong study period; code purples increased as more staff members were trained in NCI. It was vital to understand the implications of these findings to determine what benefits, if any, resulted from the training investment. To further assess the training impact, the 90-day NCI-trained staff percentage replaced the cumulative NCI-trained staff percentage as the independent variable.

Relation to Other Evidence

A related study on health care workplace violence was published by Cahill (2008). The quantitative research study by Cahill (2008) examined the impact of an 8-hr annual training regimen on nursing personnel who answered research questions to determine the following: (a) The incidence of violence in the workplace and (b) the impact of the approved training on subject and control groups as perceived by the survey respondents. The 8-hr training regimen covered the following subject matter:

  • Generational and cultural issues that impact communication styles.
  • Identification of strategies for improving communication among patients, family members, and other health care providers.
  • Behavioral characteristics associated with the cycle of aggression and violence.
  • Proper techniques for de-escalation.
  • Practice of specific evasion techniques and proper patient-containment measures.

Cahill (2008) found that caregivers in the experimental group had significantly improved mean scores on confidence in managing aggressive situations following the intervention. Cahill (2008) noted no statistical difference in the mean scores of the control group. Although the study designs differ, the findings of this QI study appear consistent with those of Cahill (2008), given the similarities of the intervention duration and nature of the training.

The code purple incidence rate of 1.46 per 1,000 ED visits compared closely with a violence incidence rate of 1.3 events per 1,000 ED visits published by Medley et al. (2012). The latter figure was derived from an ED study regarding occupancy rates and violence frequency in an ED that treats approximately 70,000 patients annually. Its violence incidence was derived from physical review of orders of emergency detainment, adverse event forms, physical restraint logs, and pharmacy records (Medley et al., 2012). A study by Kowalenko et al. (2012) exhibited incidence rates ranging from 0.18 to 2.73 per 1,000 ED visits, depending on the role of the emergency worker assaulted. A study by Brock, Gurekas, Gelinas, and Rollin (2009) yielded a mean incidence rate of 1.1 per 1,000 ED visits. The local study has a mean incidence range bracketed by the mean numbers in the studies by Medley et al. (2012) and Brock et al. (2009) as a lower bound and by the upper range in the study by Kowalenko et al. (2012) as an upper bound. The overall incidence range of the local study is also bracketed within the lower and upper bounds shared in the study by Kowalenko et al. (2012). The local study monthly incidence appears to validate code purples as a representative measure of violent events in the ED.

Limitations

Code purples represent emergency scenarios; caregivers exercise individual judgment when requesting help from the security team. The immediacy of a code purple response is consistent with other facility-wide codes such as those used for fires (code red), tornadoes (code gray), patient falls (code yellow), active firearms shooters (code silver), or internal disasters (code white). In the interest of patient and staff safety, caregivers are not required to validate code purple circumstances before engaging security on urgent matters. The possibility that an increase in code purple incidence could occur due to awareness and greater focus on violence prevention and intervention as a result of NCI training was not addressed analytically in this study. Proactive summoning of security by ED staff members was routinely practiced prior to the study; it generally occurred on a nonemergency basis, and the pattern did not appear to change subsequent to training. Subsequent studies may elect to include this consideration as a variable for explicit analysis.

More than 40% of the ED staff members were trained in NCI in the 16 months prior to the study. Any benefits derived from training that was completed before the study period were excluded when intervention benefits were analyzed (see Table 4).

Table 3
Table 3:
Cost of the intervention
Table 4
Table 4:
Code purple reductions as per the regression model

The patient population was not standardized or analyzed in terms of patient acuity, nature of affliction (trauma vs. illness), age, gender, or race. The only factor considered in the analysis was the number of psychiatric patients visiting the ED each month. Finally, the QI study was limited to the ED. Further study may determine whether similar results apply to other areas of the hospital.

Interpretation/Cost/Benefit

The correlation and regression findings reveal an interesting relationship between NCI training and code purples. The training appears to lower the incidence of code purples, some latency seems to occur before the effects of the training are visible, and the effects of the training do not appear to persist or accumulate over intervals of 6 months or longer.

Effect Timing

The delayed correlation may imply that if NCI training affects the incidence of code purple events, its effect is delayed. The periods for which negative correlations were found may imply that NCI training has a moderate-term (90–150 days) cumulative effect on the incidence of code purples if a cause-and-effect relationship can be determined by future studies. A lack of NCI training/code purple correlation for periods greater than 150 days may infer that the benefits of NCI training erode within 6 months if a cause-and-effect relationship exists. If such is the case, semiannual training may persist the negative, partial correlation between NCI training and code purple events in the ED.

NCI Training Costs

The hospital used internal trainers to deliver NCI training. The trainers were originally trained by the provider using a train-the-trainer curriculum. For the sake of completeness, the cost of training the trainers was included, even though that cost occurred before the study began. The total intervention cost of U.S. $20,008 was calculated as shown in Table 3.

Intervention Benefit

On the basis of the slope of the partial regression line, a 1% increase in NCI-trained ED staff within the current 90-day period mapped to 0.0452 fewer code purple events per 1,000 ED visits. The associated monthly NCI training impact is shown in Table 4. The total in Table 4 indicates that 33.7 code purples were avoided as a result of NCI training during the study period. The avoidance represents a 23% decrease from the projected number of code purples, which would have occurred had NCI training not taken place (based on 111 actual code purples + the total of 33.7 from Table 4).

Code Purple Cost

Unfortunately, most studies that estimate the cost of work-related violence reflect only costs associated with physical injuries, in part, because injuries can be more readily defined and measured (McGovern et al., 2000). An obvious metric for a QI study is the incidence of physical violence for health care workers published by the U.S. Department of Labor, U.S. Bureau of Labor Statistics [BLS] (2013). The latest BLS numbers from 2012 for the health care and social assistance sector reflect 15.1 violent incidents per 100,000 workers annually, with a median number of 6 workdays lost for each incident. If such a metric were applied to the study setting, one violent incident would result in U.S. $1,344 of lost productivity, equivalent to 6 days of lost wages. This amount does not include costs of medical treatment or other associated costs. However, such an event is not highly likely with a probability of 0.015%, even before one considers the 23% NCI training/code purple reduction. These figures are not definitive, as the estimated cost range varies widely. Murray (2008) conservatively estimates the cost as high as U.S. $250,000 per incident, although Murray (2008) shares very few details about the derivation of that figure.

Several additional subcomponents may contribute to the cost of a code purple incident, depending on the specific nature of the incident. Assaulted staff members may suffer short- and long-term emotional effects even if no physical injury occurs from an assault (Gates et al., 2011). The scope of assigning costs to such risks lies outside this study. Table 5 lists a number of items that code purple scenarios may encompass, even though the associated costs are not readily quantifiable.

Table 5
Table 5:
Potential cost contributors to code purple events

Cost Versus Benefit

Each code purple reduction (as per the regression model) was achieved at a cost of U.S. $593.71. This includes the original cost of training the trainers. If one considers only the cost of staff time spent on NCI training, each code purple reduction occurred at a cost of U.S. $362.26. Despite the lack of available data to support cost models for each violent event, the health care industry continues to call for efforts to reduce violence in the workplace. Ultimately, the question needed to be answered becomes: “Is investing 0.77% of annual payroll (16 of 2,080 hr per staff member) worth mitigating 23% of the workplace violence risks?”

CONCLUSIONS

Analytical focus on code purple events shall continue at the study site to optimally position security staff to respond to repeatable code purple patterns. The cost versus benefit question required informed judgment regarding continued investment on staff NCI training. For the study site, significance of the findings warranted continued investment in NCI training. Although the implied erosion of benefit outside the 150-day mark was surprising, no prior literature or studies stated or implied that benefits derived from one-time NCI training were permanent. Therefore, it may be necessary to offer biannual training to yield a further reduction in workplace violence. Subsequent studies in this field are recommended to initiate with normalized expectations regarding the implied short- and long-term effects of NCI training.

REFERENCES

Ben Naten M., Hanukayev A., Fares S. (2011). Factors affecting Israeli nurses' reports of violence perpetrated against them in the workplace: A test of the theory of planned behavior. International Journal of Nursing Practice, 17, 141–150. doi:10.1111/j.1440-172X.2011.01919.x
Brock G., Gurekas G., Gelinas A. F., Rollin K. (2009, December). Use of a “secure room” and a security guard in the management of the violent, aggressive or suicidal patient in a rural hospital: A 3-year audit. Canadian Journal of Rural Medicine, 14(1), 16–20.
Cahill D. (2008). The effect of ACT-SMART on nurses' perceived level of confidence toward managing the aggressive and violent patient. Advanced Emergency Nursing Journal, 30(3), 252–268. doi:10.1097/01.TME.0000334377.82646.04
Crisis Prevention Institute. (2013). Nonviolent crisis intervention. Retrieved from http://www.crisisprevention.com/Specialties/Nonviolent-Crisis-Intervention
Gacki-Smith J., Juarez A. M., Boyett L., Homeyer C., Robinson L., MacLean S. L. (2009). Violence against nurses working in US emergency departments. Journal of Nursing Administration, 39, 340–349. doi:10.1097/NNA.0b013e3181ae97db
Gates D. M., Gillespie G. L., Succop P. (2011). Violence against nurses and its impact on stress and productivity. Nursing Economics, 29(2), 59–67.
Gates D. M., Ross C. S., McQueen L. (2006). Violence against emergency department workers. Journal of Emergency Medicine, 31, 331–337. doi:10.1016/j.jemermed.2005.12.028
Gillespie G. L., Gates D. M., Berry P. (2013). Stressful incidents of physical violence against emergency nurses. Journal of Issues in Nursing, 18(1). doi:10.3912/OJIN.Vol18No01Man02
Inoue M., Tsukano K., Murakoka M., Kaneko F., Okamura H. (2006). Psychological impact of verbal abuse and violence by patients on nurses working in psychiatric departments. Psychiatry and Clinical Neurosciences, 60(1), 29–36. doi:10.1111/j.1440-1819.2006.01457.x
    Kowalenko T., Cunningham R., Sachs C. J., Gore R., Barata I. A., Gates D., McClain A. (2012). Workplace violence in emergency medicine: Current knowledge and future directions. Journal of Emergency Medicine, 43(3), 523–531. doi:10.1016/j.jemermed.2012.02.056
    Leonard K. L. (2008). Is patient satisfaction sensitive to changes in the quality of care? An exploitation of the Hawthorne effect. Journal of Health Economics, 27, 444–459. doi:10.1016/j.jhealeco.2007.07.004
    Luck L., Jackson D., Usher K. (2008). Innocent or culpable? Meanings that emergency department nurses ascribe to individual acts of violence. Journal of Clinical Nursing, 17, 1071–1078. doi:10.1111/j.1365-2702.2006.01870.x
    Mayo E. (1933). The human problems of an industrial civilization. New York, NY: MacMillan.
    McGovern P., Kochevar L., Lohman W., Zaidman B., Gerberich S. G., Nyman J., Findorff-Dennis M. (2000). The cost of work-related physical assaults in Minnesota. Health Services Research, 35, 663–686.
    Medley D. P., Morris J. E., Stone C. K., Song J., Delmas T., Thakrar K. (2012). An association between occupancy rates in the emergency department and rates of violence toward staff. Journal of Emergency Medicine, 43, 736–744. doi:10.1016/j.jemermed.2011.06.131
    Murray J. S. (2008). No more nurse abuse. American Nurse Today, 3(7). Retrieved from http://www.americannursetoday.com/ContentList.aspx?fid=4598
    National Institute for Occupational Safety and Health. (2002). Violence occupational hazards in hospitals. Retrieved from http://www.cdc.gov/niosh/docs/2002-101
      Pai H. C., Lee S. (2011). Risk factors for workplace violence in clinical registered nurses in Taiwan. Journal of Clinical Nursing, 20(9–10), 1405–1412. doi:10.1111/j.1365-2702.2010.03650.x
        Sato K., Wakabayashi T., Kiyoshi-Teo H., Fukahori H. (2013). Factors associated with nurses' reporting of patients' aggressive behavior: A cross-sectional survey. International Journal of Nursing Studies, 50(10), 1368–1376. doi:10.1016/j.ijnurstu.2012.12.011
        U.S. Department of Labor. (2013). Nonfatal occupational injuries and illnesses requiring days away from work, 2012 (USDL-13-2257). Washington, DC: Government Printing Office.
          Winstanley S., Whittington R. (2004). Aggression towards health care staff in a UK general hospital: Variation among professions and departments. Journal of Clinical Nursing, 13, 3–10. doi:10.1111/j.1365-2702.2004.00807.x
            Keywords:

            hospital violence; nursing violence; violence de-escalation; violence incidence; violence training; workplace violence

            © 2014 by Lippincott Williams & Wilkins.