Nurse-Reported Workplace Violence: Results from a Repeated Statewide Survey

Abstract

Background:

Workplace violence significantly affects registered nurses, contributing to burnout and intention to leave.

Methods:

The Michigan Nurses Study conducted surveys in 2022 and 2023, examining the prevalence of verbal, physical, and sexual violence, and coworker bullying. Personal and workplace factors associated with reporting any violent event were examined using multivariable logistic regression.

Results:

There was a decline in overall workplace violence from 50.2% to 43.4%, despite a rise in sexual harassment from 9.9% to 11.8%. Over half of the events were patient sourced. Underreporting remained a problem. Factors associated with increased likelihood of reporting any violent event included younger age, employment in acute care and long-term care settings, unfavorable practice environments, less confidence in management, inadequate support for workplace stress, and understaffing on the past shift.

Conclusions:

Despite a decrease in reported violent events, incidents remain common. Improvements in working conditions, management of understaffing, and support systems are crucial to mitigate workplace violence against nurses.

INTRODUCTION

Workplace violence among the registered nurse workforce comprises a substantial occupational threat in the United States. Relative to other employment sectors, health care settings report high annualized violent incident rates (Hawkins & Ghaziri, 2022). As an occupational category, nurses rank highly among those who experience workplace violence (National Institute for Occupational Safety and Health, 2022). These events are not innocuous; workplace violence is associated with nurses’ increased intention to leave the workplace, increased thoughts of self-harm, and lower overall well-being scores (Titler et al., 2024). Such events also contribute to workforce instability, an estimated 13% of nurses who experience workplace violence miss workdays as a result of the event (American Nurses Association, 2018).

While prevalence estimates vary, prior research suggests that nurses, as an occupational category, face substantial risks, with up to 1 in 4 nurses experiencing workplace violence (American Nurses Association, 2015). A systematic review and meta-analysis of 65 studies estimated a pooled prevalence of physical violence against nurses at 19.3%, with slightly higher prevalence estimated for nurses (22.9%), and studies based in North America (23.6%) (Li et al., 2020). Systematic reviews have characterized antecedent factors for nurses’ workplace violence, including both nurses’ personal characteristics (female gender, younger age) and workplace factors (staffing levels, quality of the working conditions) (Nowrouzi-Kia et al., 2019).

Despite a robust literature base, there remain important knowledge gaps. First, studies often focus on one type of violence (primarily patient-to-nurse physical events), while ignoring verbal threats or staff-on-staff events, which prohibits a holistic assessment of workplace safety. Secondly, many studies do not address whether the nurse reported the incident to their employer and critically, the assessment of the employer’s response. Third, sample sizes in the reviewed literature are small; of the 65 studies included in the meta-analysis, ten were from the United States, and only six focused on nurses, with the sample sizes reported ranging from 55 to 3,999 participants (Li et al., 2020). The systematic review of antecedent factors of workplace violence against nurses included 13 studies, five of which were US-based with a range of sample sizes from six to 3,465 (Nowrouzi-Kia et al., 2019). Many studies focus solely on the characteristics of the nurse or perpetrator and do not consider potentially important organizational factors that may be associated with violent events. As opposed to personal characteristics, there may be added possibilities to reduce violent events through changes in modifiable workplace characteristics. These limitations in the literature have hampered the development and testing of effective interventions.

In this context, to narrow knowledge gaps and inform future institutional and policy interventions, we conducted a secondary analysis of data collected from the Michigan Nurses Study to gauge practicing nurses’ experiences with workplace violence at two time points: 2022 and 2023 (Medvec, Marriott, et al., 2023). The 2022 survey was a larger effort to gauge perspectives of Michigan nurses – both those in practice and those not in practice – on key health care delivery issues. Given the high rates of violent events reported, we added additional survey questions in 2023 to gauge the experiences of practicing nurses and to characterize the source(s) of violence (patient, visitor, another team member) and nurses’ experiences in reporting incidents to their employer. The goal of this effort was to identify patterns in workplace violence, characterize violent events more thoroughly, and identify factors that contribute to an increased risk of workplace violence.

METHODS

Population.

The Michigan Nurses’ Study is an email-based survey sent to registered nurses in Spring 2022 and repeated in Spring 2023. The survey methods have been published previously (Medvec, Marriott, et al., 2023; Medvec, Titler, et al., 2023). For each study year, individuals who held an unrestricted registered nurse license with a valid email address recorded by the Michigan Board of Nursing were eligible to participate. Individuals with restricted licenses or noted as in a disciplinary process were excluded, as were individuals who did not have individual, valid email addresses on file with the state.

Qualtrics^™ (Provo, UT) was used for both survey distribution and management. The study team made up to three email attempts to reach potential participants, eight days apart. In the introductory materials, potential participants were invited to receive electronic copies of study findings upon completion. The 2022 survey was open from February to March and the 2023 survey ran from May to June.

All participants completed online consent. At two points in the survey, all participants received contact information for state and federal mental health assistance. The University of Michigan Institutional Review Board reviewed the protocol and determined it to be exempt from ongoing IRB review. Participants were asked to respond to all survey questions regarding their primary nursing position and could skip any questions they wished. STROBE criteria for cross-sectional surveys were used to organize both the overall project and reported findings (von Elm et al., 2014).

Study Measures.

Primary Outcome.

The primary outcome was a composite measure of reporting any workplace violence – verbal, physical, or sexual – in the past year. Specific questions are included in the Appendix. Operational definitions from the World Health Organization were provided to participants and each outcome was dichotomized (yes/no) (World Health Organization, 2003). Subset analyses were performed to examine each of the three events separately.

Violent Events Context.

Participants who reported any violent event were asked the source (patient, family, visitor, fellow health care worker), whether they reported the event to their manager (yes/no), whether they received an employer response (yes/no), and if so, whether they were satisfied with the response (yes/no).

Individual Covariates.

Age was treated as a categorical variable, with pre-defined cut points that aligned with Michigan’s nurse licensure reporting system (< 25 years, 25–34 years, 35–44 years, 45–54 years, 55–64 years, ≥ 65 years). Participants self-reported their race and ethnicity and could choose multiple identities (Asian, Black of African American, Hispanic or Latino, Middle Eastern or North African, Native American or Alaska Native, Native Hawaiian or Other Pacific Islander, White, Prefer Another Answer, or Unknown). We also asked nurses to report their gender identity (female, male, transgender, nonbinary, gender conforming, or prefer another choice). To protect participant privacy, we collapsed responses from individuals who identified as one of the four latter categories.

Workplace Covariates.

Informed by the literature and our conceptual model, we posited that several workplace factors would be associated with violent event reports. These included primary practice setting (e.g., inpatient, long-term care, community/public health, school nursing, and education), and individual questions on the quality of their practice environment (unfavorable, mixed, favorable), (Friese, 2012) confidence in management to resolve problems (4-point scale, very confident to not at all confident), (Aiken et al., 2023) provision of adequate support for workplace stress (5-point scale, strongly agree to strongly disagree) and characterization of staffing on the last shift worked (understaffed, staffed appropriately, or overstaffed) (Kalisch et al., 2011).

Analyses.

SAS 9.4 was used for all analyses (Cary, NC). We used descriptive statistics and calculated 95% confidence intervals to compare violent event rates reported in the 2022 survey with the 2023 survey and examined missingness of all dependent and independent variables. Next, using data solely from the 2023 cohort, we used bivariate analyses and multivariable logistic regression to examine the relationship between the reported violent events and workplace and individual factors. Model assumptions were checked using standard techniques, including residual plots and goodness of fit tests. Variable selection for the multivariable model was pre-specified by our prior work, (Medvec, Marriott, et al., 2023; Titler et al., 2024), the National Academy of Medicine conceptual model for clinician well-being (Brigham et al., 2018), as well as extant theory in the health care worker violence literature (Berger et al., 2024; Cheng et al., 2020; Findorff, 2004; C. B. Jones et al., 2023; M. Jones, 2021; Kim et al., 2023; Sari et al., 2023). Specifically, prior work that has examined workplace violence among nurses has identified age, gender, staffing characteristics, and workplace environment as key antecedents (Nowrouzi-Kia et al., 2019).

RESULTS

In 2022, 13,687 nurses completed the survey from 164,903 delivered surveys (8.3% overall response rate). In 2023, 10,277 nurses completed the survey from 138,455 delivered surveys (7.4% overall response rate). Because of our interest in understanding violent events experienced by practicing registered nurses, the analyses below include the subsamples of 7,983 nurses in 2022 and 6,545 nurses in 2023 who were in active clinical practice and responded to the workplace violence questions. The primary analysis used complete case analysis. The appendix table summarizes the individual participant characteristics by year and compares them with available demographic data from the Michigan Board of Nursing. The appendix also includes a sensitivity analysis of the multivariable model using multiple imputation, with no appreciable differences in the results reported herein.

Violent Event Reports and Context in 2022 and 2023.

Figure 1 compares the distribution of overall, verbal, physical, and sexual violence events reported by nurse participants. The proportion of nurses (with corresponding 95% confidence interval) who reported any event declined from 2022 to 2023, from 50.2% (95% CI 49.2–51.2) to 43.4% (95% CI 42.2–44.6), respectively. While both verbal and physical events declined, the number of nurses who reported workplace sexual harassment increased from 9.9% (95% CI 9.3–10.5) in 2022 to 11.8% (95% CI 11.0–12.6) in 2023. These differences and their 95% confidence intervals do not overlap, suggesting notable differences between 2022 and 2023.

Figure 1.

Comparison of Nurse-Reported Workplace Violence in 2022 and 2023

Next, we examined whether nurses reported events to their employer and perceived employer response, as summarized in Table 1. In 2022, 52.4% of nurses (95% CI 52.8–55.6) compared with 43.1% (95% CI 41.4–44.9) in 2023 did not report events to their employer. Of those who did, employers responded less than half of the time. The majority of those who received an employer response were satisfied with the outcome (66.4% and 71.2%, respectively).

Table 1.

Nurse Reports of Workplace Violent Events to their Employer and Employer Response

	2022 n=4,591		2023 n=3,063
	n(%)	95% CI	n(%)	95% CI
Reported event to employer
-Yes	2,061 (44.9)	43.5–46.3	1,695 (55.3)	53.5–57.1
-No	2,489 (54.2)	52.8–55.6	1,319 (43.1)	41.4–44.9
-No Response	41 (0.9)	0.7–1.2	49 (1.6)	1.2–2.0
Employer responded to report
-Yes	887 (43.0)	41.6–44.4	650 (38.4)	36.7–40.1
-No	1,146 (55.6)	54.2–57.0	1,039 (61.3)	59.6–63.0
-No Response	28 (1.4)	0.03–0.3	6 (0.4)	0.2–0.6
Nurse satisfied with employer response
-Yes	589 (66.4)	65.0–67.8	463 (71.2)	69.6–72.8
-No	291 (32.8)	31.4–34.2	185 (28.5)	26.9–30.1
-No Response	7 (0.8)	0.5–1.1	2 (0.3)	0.1–0.5

Violent Event Report Context in 2023

In the 2023 nurse survey sample, the primary source of violence was patients (55.5%), followed by another health care worker (26.1%), and patients’ families and visitors (18.8%) (Table 2). Among health care workers, the sources of violence were managers (9.4%), fellow nurses (6.8%), other health care team members (5.2%), and physicians (4.7%).

Table 2.

Source of Workplace Violence, 2023.

	n	%
Patients	1669	54.5
Patients’ Families and Visitors	577	18.8
Managers	288	9.4
Nurses	208	6.8
Other members of the health care team	159	5.2
Physicians	145	4.7
No Response	17	0.6

Factors Associated with Violent Event Reports in 2023

The bivariate distribution of violent event reports by selected individual and workplace factors are shown in the online appendix. In multivariable analyses, the following factors were associated with the likelihood of reporting a violent workplace event in the past year (Table 3): nurses below the age of 45 were significantly more likely to report events, whereas nurses over 55 were significantly less likely to report events. Compared with males, females were less likely to report events (Odds Ratio (OR) 0.74, 95% Confidence Interval (CI) 0.60–0.91). Compared with nurses who reported white race, Asian nurses and Black or African American nurses were significantly less likely to report events.

Table 3.

Factors Associated with Reporting Any Workplace Violence, 2023. N=5,439

	Odds Ratio	95% CI	Standard Error
Individual Factors
Age
- < 25 years	1.69	(1.13, 2.51)	0.20
- 25–34 years	1.73	(1.41, 2.11)	0.10
- 35–44 years	1.27	(1.06, 1.51)	0.09
- 45–54 years	REF	REF	REF
- 55–64 years	0.80	(0.68, 0.95)	0.08
- ≥ 65 years	0.41	(0.32, 0.52)	0.12
Self-Reported Gender Identity
- Female	0.74	(0.60, 0.91)	0.11
- Male	REF	REF	REF
- Reported another identitya	1.41	(0.49, 4.06)	0.54
Race and Ethnicityb
- Asian	0.46	(0.30, 0.70)	0.22
- White	REF	REF	REF
- Black or African American	0.68	(0.5, 0.94)	0.16
- Hispanic or Latino	0.73	(0.44, 1.21)	0.26
- Middle Eastern/Northern African	0.74	(0.35, 1.55)	0.38
- Native American, Alaska Native, Native Hawaiian, Pacific Islander	1.80	(0.7, 4.68)	0.49
- Reported multiple identities	1.22	(0.86, 1.74)	0.18
- Other Answers	1.29	(0.81, 2.05)	0.24
- Unknown	1.05	(0.41, 2.72)	0.48
Workplace Factors
Setting
- Inpatient/Acute Care	1.76	(1.45, 2.15)	0.10
- Long-Term Care	1.28	(0.95, 1.74)	0.16
- Community/Public Health	REF	REF	REF
- School Nursing	1.29	(0.81, 2.05)	0.24
- Nursing Education	0.83	(0.52, 1.34)	0.24
- Other Practice Setting	0.92	(0.75, 1.13)	0.10
Practice Environment Assessment
- Unfavorable	3.21	(2.49, 4.16)	0.13
- Mixed	1.60	(1.37, 1.86)	0.08
- Favorable	REF	REF	REF
Confidence in Management
- Somewhat/Not at All Confident	1.43	(1.23, 1.66)	0.08
- Confident/Very Confident	REF	REF	REF
Adequate Support for Workplace Stress
- Somewhat/Strongly Disagree	1.64	(1.39, 1.92)	0.08
- Neutral	1.07	(0.91, 1.27)	0.09
- Somewhat/Strongly Agree	REF	REF	REF
Staffing on Last Shift
- Understaffed	1.43	(1.25, 1.63)	0.07
- Staffed Appropriately	REF	REF	REF
- Overstaffed	1.63	(0.96, 2.77)	0.27

^aOther identity choices included transgender, nonbinary, gender conforming, or prefer another choice

^bParticipants could choose multiple identities.

Workplace factors were also significantly associated with violent events. These included inpatient/acute care setting (versus community or public health), where nurses were 76% more likely to report an event (OR 1.76, 95% CI 1.45–2.15). Practicing in an unfavorable (OR 3.21, 95% CI 2.49–4.16) or mixed (OR 1.60, 95% CI 1.37–1.86) practice environment, compared with a favorable environment was associated with significantly higher likelihood of reporting events. Event reports were associated with less confidence in management to resolve problems (OR 1.43, 95% CI 1.23–1.66), inadequate support for workplace stress (OR 1.64, 95% CI 1.39–1.92), and understaffing on the last shift worked (OR 1.43, 95% CI 1.25–1.63). In a sensitivity analysis of the multivariable model using multiple imputation to account for missing data, results were not notably different from the results reported herein.

DISCUSSION

In this repeated cross-sectional survey to registered nurses in Michigan, workplace violence occurred frequently, yet these events were underreported to employers, and varied by notable personal and workplace factors. Nearly half of all surveyed nurses in both years experienced some form of workplace violence. This suggests that workplaces do not provide a consistently safe atmosphere to facilitate effective nursing care delivery. While fewer nurses reported events to their employer in 2023 than in 2022, the prevalence of non-reporting exceeds 40% for both years. Also concerning is that of those who reported, employers responded less than half of the time. Given our team’s prior work to establish a relationship between workplace violence and intention to leave the workplace (Medvec, Marriott, et al., 2023), employers who seek to retain existing personnel should (1) respond meaningfully when nurses report violent events and (2) employ interventions to address root causes of workplace violence that are directed towards registered nurses. Given the high frequency of verbal events, prevention and early mitigation of verbal violence is an important but under addressed target for management response and clinician education.

Employer response to workplace violence is an important contributor to job satisfaction. The substantial proportion of nurses in both years who stated their employer took no action on reported events is a missed opportunity to improve employee safety and well-being. An encouraging note is that among nurses who reported and received an employer response, most were satisfied with the outcome. More troubling is that nearly half of the sample did not report the event to their employer. This high rate of underreporting could indicate that nurses presume that events are inherent to the workplace, they are fearful of negative consequences, that the event did not affect them negatively, or that they do not expect their employer to respond in a meaningful way (Kim et al., 2023). However, the underreporting of workplace violence restricts opportunities for quality improvement and reaffirms the importance of leadership efforts to promote workplace psychological safety. A 2021 scoping review that included 49 studies identified underreporting as a major concern, yet violent event reports have increased steadily over time, suggesting that nurses are increasingly willing to report these concerns (Huang et al., 2022). Opportunities to improve the quality of violent event reporting include increased awareness of how workplace violence negatively impacts health care workers, that the environment will not improve without consistent and reliable reporting, and that leaders should take workplace violence reports seriously and support affected nurses. Further, leaders must implement organizational interventions to stop violent events and engage with nurses to create a culture where nurses feel psychologically safe to report such events.

The findings also suggest notable differences in workplace violence by nurse-specific and workplace factors. Nurses below the age of 44 report higher rates of events, whereas nurses over the age of 55 report fewer events. Women in our sample are less likely to report events compared with men. Nurses who identify as a member of the gender minority community report higher rates of events, but these results were not statistically significant. The latter finding likely reflects a relatively low number of individuals who chose a gender minority category. Despite documented concerns for racism inflected among minority nurses, Asian and Black nurses were less likely to report events than white nurses. The differences observed by age, gender, and race/ethnicity may reflect underlying response patterns and not necessarily actual differential rates of violent event occurrences.

Workplace factors were highly associated with violence reports. Acute care nurses reported more events, as did nurses who, regardless of their clinical setting, expressed less confidence in management, cited inadequate support for workplace stress, or reported understaffing on their last shift. These findings suggest that there are actionable targets for health care systems to implement in their operating plans to support their clinical workforce. These include improved communication strategies between clinicians and managers, correction of chronic understaffing, and provision of adequate employee support services. Given our findings of increased violent events among younger employees, special attention should be paid to assuring that supportive services meet the needs of the newest entrants.

The July 2023 American Academy of Nursing policy dialogue on workplace violence generated multiple strategies, including improved training, event reporting and analysis, collaborative efforts to share best practices, and resource provision to implement promising prevention and mitigation efforts (American Academy of Nursing, 2024). There are also state and federal policy approaches to addressing health care workplace violence. In December 2023, Michigan’s Governor Gretchen Whitmer signed bipartisan legislation that increases fines for health care facility visitors who verbally or physically assault health care personnel (Ninan et al., 2024). The bill, as enacted, would apply to only 19% of the events reported by study participants. The bill also requires health care facilities to post prominent signage that reflects the increased fees. Other anti-violence measures reported in the literature, including required annual anti-violence plans, compulsory reporting of violent events, and other penalties, were not included in the legislation. Specifically, increased fines or other enforcement efforts against violent events instigated by patients was excluded from the legislation (the source of 54.5% of the reported incidents), due to concerns for extraneous law enforcement involvement in clinical matters and potential negative impacts on patients from historically marginalized groups (Rayasam, 2023). Policy approaches to address health care workplace violence will require a careful balance of protecting individual autonomy and creating safer environments for clinicians to deliver care.

Regulators have an important role to play to assure health care worker safety. In 2021, the Joint Commission added three new elements of performance and revised two existing elements to promote health care worker safety (The Joint Commission, 2001). These changes require institutions to implement a new annual worksite analysis to identify violence risks, offer enhanced health care worker training, and establish a multidisciplinary workplace violence prevention program. Institutional performance on these elements, consequences of nonperformance, and the impact of these revisions is not yet clear.

Limitations

A key limitation to the project is the low response rates obtained in both 2022 and 2023. The low response rates pose risks to the internal validity of the project. While we did not identify major demographic differences between responders and non-responders and the survey introductory material made no mention that workplace violence questions were included, it is possible that respondents to a confidential survey may hold more negative views of the workplace and thus, inflate the prevalence estimates reported. Asking participants to report events that occurred over a one-year time period may be subject to recall bias. The selection of binary outcomes, instead of frequency counts, is another contextual limitation.

Because personal identifiers were not linked to survey results to protect privacy, we are unable to identify participants who completed surveys in both 2022 and 2023, which prohibits multivariable analyses with both years of data included. There is likely overlap, but when testing differences over time, we assumed the samples were independent. Few participants from groups that are historically excluded from research - across race, ethnicity, and gender identity domains - limit the interpretation of our findings to these important and understudied populations. The reporting patterns observed by gender, race, and ethnicity may reflect underlying differences in survey reporting, rather than actual differences in events that occurred.

Future study samples enriched with these individuals that combine quantitative and qualitative approaches would facilitate exploration of the differences reported herein and apply extant intersectional approaches to understanding this issue. Despite these limitations, strengths of the study include the statewide sampling approach to account for differences across clinical settings, the use of previously used and validated measures, and the ability to compare differences over a recent time period.

Implications and Conclusions

A high proportion of practicing registered nurses in Michigan who responded to a voluntary survey reported workplace violence, which threatens the ability for health care systems to deliver safe and effective care. Workplace violence correlates with worker mental health, burnout, and intention to leave the workforce. Given the pervasive challenges that threaten the US health care workforce, increased attention should be paid to understanding and addressing the root causes of clinician departures. Among these, detection and prevention of workplace violence is ripe for improvement. Institutional, state, and federal policy options are available to address these concerns and build the infrastructure necessary to understand, prevent, and ameliorate this large and potentially growing concern.

Registered nurse-reported workplace violence can be considered an occupational injury. Such widespread occurrences with negative consequences compel organizational change. Applying the National Institute for Occupational Safety and Health (NIOSH) hierarchy of controls framework for workplace injury risk reduction, the first (and likely most effective) control strategy is to eliminate the hazard, with the least effective strategy identified as personal protective equipment (National Institute for Occupational Safety and Health, 2023). Therefore, to eliminate the hazard of nurses’ workplace violence, strategies include correcting chronic understaffing, providing additional staff support, strengthening staff and leadership relationships, which collectively will improve nurses’ practice environments. The current study findings align with recommendations from the NIOSH workplace injury prevention framework, which focuses on organizational-level interventions. Unless organizational improvements have been enacted and are evaluated, downstream interventions are unlikely to be effective.

Our findings have implications for practicing nurses, health care executives, and policymakers. Nurses should be empowered to freely report violent events to their employers, without fear of negative consequences. Without detailed and standardized reporting, prevention and mitigation strategies will be incomplete and ineffective. Huang et al (2022) recommend implementing violent reporting relationships outside of the direct chain of command, such as to occupational health or a safety officer instead of a direct supervisor. Executives should first encourage their workforce to report events and act earnestly after reports are filed. Optimal report response includes supporting the individual who experienced the event and advocating for system-level changes to address root causes. Finally, policymakers must recognize that nurses and other health care workers routinely experience workplace violence. Legislative and regulatory proposals to reduce the burden of health care workplace violence need to address the underlying causes and strengthen event reporting.