A labor perspective of workplace violence prevention: Identifying research needs1

Introduction 

Labor unions want safe environments in which workers can perform their duties in a supportive atmosphere and return home at the end of a shift without suffering injury or illness. Workplace violence and the lack of programs to prevent it all too often stand in the way of this fundamental union goal.

Fatalities 

In 1998, the U.S. Bureau of Labor Statistics (BLS) reported 709 fatalities as a result of job-related homicide, making it the second leading cause of death on the job.1 The 1998 numbers reflect a significant decrease, especially in retail trade in which homicides fell by 46% from 1994. Although this reduction is impressive, the risk of death on the job in our nation’s retail, taxicab, and police agencies continues to be a significant problem that demands a studied and diligent response. Twenty-two million workers are employed in retail trade,2 an industry that accounted for 40% of all workplace homicides. Table 1 shows the decline in violence-related fatalities in select industries between 1994 and 1998. From 1997, the 18% decline in workplace homicides was significantly greater than the overall 8% drop in homicides reported by the U.S. Justice Department. A total of 223 on-the-job suicides occurred in 1998, up from 216 the previous year.

Table 1. Number of occupational homicides by industry, 1994–1998legend

	Industry	1994	1995	1996	1997	1998
	Total homicides	1080	1036	927	860	709
	Retail trade	530	422	437	395	286
	Grocery stores	196	152	146	141	95
	Eating and drinking establishments	135	121	135	109	69
	Gasoline service stations	41	36	23	34	24
	Taxicab	87	68	50	74	48
	Detective and armored-car services	49	27	29	21	18
	Police protection	65	61	45	61	50

legend Source: Bureau of Labor Statistics.1

For female workers in 1998, homicide was the leading cause of death on the job: 168 out of 482 fatalities (34%) an increase from the 145 homicides in 1997.

Unlike homicide, on-the-job suicide has not declined. Could work-related stress be increasing in conjunction with economic restructuring, for example, with increases in low-wage service-sector jobs (11 million new jobs added in the past decade), lower rates of unionization, decreased health care benefits, stagnant wages, job instability in connection with downsizing, de-industrialization, and the proliferation of temporary work (11.8 million jobs added between 1988 and 1998)? Other organizational factors such as swing shifts, mandatory overtime, increased pace of work, self-managed work teams, and quality management systems may also factor into increased job stress.3

Nonfatal assaults 

Data from the National Crime Victimization Survey for 1992 to 1996 indicate that 2 million nonfatal assaults occur in the workplace each year.4 Figure 1 lists select occupations with a large number of victimizations.

View Large Image Download to PowerPoint

Figure 1. Occupations at high risk of violence and the average number of victimizations in the workplace between 1992 and 1996. Source: U.S. Bureau of Justice Crime Victimization Survey.4

In 1996, nonfatal violence resulted in 876,000 lost workdays and $16 million in lost wages. A disproportionate number of nonfatal assaults involved public-sector workers who made up 16% of the workforce in 1996, but made up 37% of the victims (Table 2). Public service jobs include frequent contact with the public, and changes in public policy have led to cuts in social services and health care, exacerbating already tense working environments. Social service, criminal justice, and public-sector health care agencies serve populations at the bottom of the socioeconomic ladder and who are most affected by society’s social ills, including unemployment, alcohol and drug abuse, mental illness, disease, crime, and violence.5

Table 2. Employers of workplace violence victims, 1992–1996legend

	Nonfatal assaults	Numbera	%
	Private employer	1,127,800	56.1
	State/local government	677,600	33.7
	Self-employed	134,500	6.7
	Federal government	62,900	3.1
	Not ascertained	35,100	0.3

a Annual average, 1992–1996. legend Source: U.S. Bureau of Justice, Crime Victimization Survey.4

The health care industry has experienced especially high rates of violence. The 1997 BLS data on nonfatal assaults that cause lost time from work revealed the following:

Public schoolteachers also have experienced a disturbingly high rate of assault, especially in junior high and special education. Police, county, and state-run health care institutions have high rates as well.

On some psychiatric units, assault rates are more than 100 cases per 100 workers per year.6 A Washington State Department of Labor and Industries study of workers’ compensation claims for 1992 to 1997 showed that psychiatric hospitals had the highest average rate of claims related to workplace violence of all industries: 872 per 10,000 workers. The second, third, and fourth highest claim rates also occurred in health care as follows: residential care, 417/10,000 workers; skilled nursing care, 254/10,000 workers; and nursing/personal care, 240/10,000 workers. These rates compare with a rate of 19/10,000 workers for all industries.7

In the private sector, retail workers have a high level of contact with the public, which has unrestricted access to the workplace. Exchanging money with the public and the fact that many supermarkets and convenience stores are open 24 hours contribute to exposure to violence.

Table 3 shows rates of nonfatal violence by occupation. Although considerable research has provided information on rates of nonfatal violence by occupation (BLS, Washington State, Bureau of Justice), significant differences exist in the techniques used and the databases accessed. This results in wide variation in estimated rates of nonfatal violence.

Table 3. Occupations of victims of nonfatal workplace violence, 1992–1996legend, legend

	Occupation	Average	Annual rate/1000 workers
	Totala	2,009,400	14.8
	Medical
	Physicians	10,000	15.7
	Nurses	69,500	24.8
	Technicians	24,500	21.4
	Other	56,800	10.7
	Mental health
	Professional	50,300	79.5
	Custodial	8,700	63.3
	Other	43,500	64.0
	Teaching
	Preschoolb	2,400	3.6
	Elementary	35,400	16.0
	Junior high	47,300	57.4
	High school	33,300	28.9
	College/university	6,600	2.5
	Technical/industrialb	400	4.4
	Special education	9,000	40.7
	Other	14,400	10.1
	Law enforcement
	Police	234,200	306.0
	Private security	71,100	117.3
	Corrections officers	58,300	217.8
	Other	67,600	61.5
	Retail sales
	Convenience/liquor store	61,500	68.4
	Gas station	15,500	79.1
	Bar	26,400	91.3
	Other	228,200	17.5
	Transportation
	Taxi drivers	16,100	183.8
	Bus drivers	17,200	45.0
	Other	43,200	10.0
	Other/unspecified	758,000	8.2

legend Rates are calculated using population estimates from the National Crime Victimization Survey for occupation, 1992–1996. Detail may not add to total because of rounding. a The total for specified occupations was 1,251,400, with 29.4 victims of workplace violence per 1,000 workers. b Fewer than 10 sample cases. legend Source: U.S. Bureau of Justice, Crime Victimization Survey.4

Worker-on-worker violence and profiling 

The media, employers, and consultants have given a disproportionate amount of attention to worker-on-worker violence. Sensational media coverage has inspired this focus. The BLS 1998 Census of Fatal Occupational Injuries database shows that coworkers or former coworkers perpetrated 9% of workplace homicides, compared with 79% who committed robberies and other crimes. The BLS 1997 report on nonfatal injuries indicates that coworkers committed assaults resulting in lost time in 7% of cases compared with 55% of cases committed by health care patients.

A study by the National Institute for Occupational Safety and Health (NIOSH) and the U.S. Postal Service revealed that the rate of violence in the postal service, 0.63 per 100,000 workers, is similar to the rate for all industries of 0.64 per 100,000.8 This report concluded that “Neither the Postal Service industry nor postal occupations are among the groups at increased risk for work-related homicide.” These facts fly in the face of the popular misconception about “going postal” and the singular emphasis on coworker violence. This has resulted in a proliferation of management consultants focused on “worker profiling” to screen out workers considered to have a potential for violence.

Psychologist Mark Braverman9 wrote, “Focusing violence prevention efforts on prediction based on ‘profile’ list leads to panicky, and often disastrous ‘witch hunt’ reactions. With respect to pre-employment screening it raises troublesome legal issues.” Denenberg et al.10 emphasize the importance of building systems to prevent workplace violence, rather than focusing on behavior of individuals. They stress the importance of organizations becoming “violence prepared” rather than “violence prone.” This logic is consistent with Occupational Safety and Health Administration (OSHA) Violence Prevention Guidelines that emphasize a participative, programmatic approach. Dealing with interpersonal conflict in the workplace is important. However, in contrast to the profiling approach, a systems approach allows for identification and analysis of and action on multiple stress factors.

Zero tolerance 

“Zero Tolerance for Violence” is a popular slogan included in many newly written policies on workplace violence that reflects a worker-based approach. Rather than addressing organizational stress factors that may lead to conflicts, this approach focuses on weeding out people who reach the breaking point and not on reducing stress before people reach that point.

Denenberg and Braverman9, 10 make a strong case that such policies may lead to violating “just cause” standards if management discharges combatants without carefully weighing facts and circumstances and proving the penalty fits the offense. Workers may view these policies as unfair if policies are arbitrary and reflexive.

Federal OSHA guidelines 

In 1993, NIOSH published an alert to encourage action to prevent workplace homicide.11 Detailed research that revealed the significant role that violence plays in our society’s overall burden of occupational morbidity and mortality followed this alert.12 The NIOSH research, combined with increased advocacy by labor unions and public health forces, influenced OSHA to develop guidelines.13, 14 The OSHA guidelines for health care and social service workers were based largely on work done by the California OSHA and the State of Washington Department of Labor and Industries. The guidelines include the following elements: management commitment and employee involvement, a written program, a work site analysis, hazard prevention and control, medical management and post incident response, training and education, and record keeping and evaluation of the program. The guidelines also include sample surveys, checklists, and report forms to simplify risk analysis and program implementation. In fact, during the past 4 years, a proliferation has occurred of similar guidelines developed by governmental, labor, and private groups.15, 16, 17, 18, 19, 20, 21, 22 Labor unions have used these materials to encourage development of voluntary programs.

Voluntary implementation of federal OSHA recommendations and guidelines 

Although publication of OSHA guidelines has inspired some managers to act, the sparse research on this subject shows that few employers are implementing full programs. A recent survey by the American Society of Safety Engineers (ASSE) of West Virginia schools revealed that most are unprepared to deal with school violence. The survey assessed whether appropriate emergency procedures were in place and whether teachers and administrators were knowledgeable and prepared to implement response plans.23 Studies have consistently found that retail businesses face an elevated risk of robbery during the night hours,24, 25 when staffing is at the minimum. Many retail employers simply respond by providing more security, such as installing metal detectors or hiring a security person for the night hours.

A survey conducted in 1998 of 1000 risk managers and ASSE members showed that 70% worked in settings with no formal risk assessments for workplace violence.26 The survey also revealed that employers were more likely to train managers to identify signs of violent behavior among employees than to provide employee training on conflict resolution.

Labor finds it particularly problematic that no OSHA standard exists on workplace violence, despite the fact that workplace violence is a significant cause of death and injury. Enforceable standards provide a strong incentive for employers to work with unions to implement safety programs. Such standards also provide workers and unions with recourse if employers are not interested in labor/management cooperation and refuse to implement minimum protections. Arguments against a standard include the notion that employers will voluntarily implement programs. This is contrary to labor’s experience as well as to the earlier cited surveys.

The dramatic decline in hepatitis B among health care workers that took place after OSHA enacted the “occupational exposure to bloodborne pathogens” standard exemplifies the powerful impact an OSHA standard can have. Voluntary programs were inadequate despite years of widespread dissemination of guidelines by the Centers for Disease Control and Prevention (CDC), industry, and labor groups that encouraged hepatitis B immunization and training programs for health care workers. Before the vaccine’s introduction, the CDC estimated that 6000 to 8000 health care workers were infected with hepatitis B each year, and 200 to 300 died annually. Today, the CDC estimates the number of new cases of hepatitis B among health care workers has fallen to 400 per year.27 Although regulation has proven a powerful method for controlling workplace hazards, enacting new rules in today’s anti-regulatory environment is particularly difficult.

Short staffing 

Workers often raise the issue of short staffing as an important factor in workplace violence. In health care, for example, a staffing pattern of three workers for a ward of 28 psychiatric patients is quite common. Workers have little time to pay individual attention to patient needs in these circumstances, as tasks such as serving meals, transporting patients to programs, distributing medications, and documentation consume their time. Other stress issues related to staffing include erratic shift patterns, mandatory overtime, fatigue, effects of work on family life, and health effects such as hypertension.9

Employers, OSHA, and researchers alike are reluctant to address staffing issues. Employers use the management rights argument: the notion that management has a supreme right to manage and therefore determine staffing levels.

The state OSHA plan in New York for public employees (PESH) has issued several general duty clausea violations that concern workplace violence. However, none of them cites staffing even though the complaints listed it as a factor and the affected unions considered staffing an important element. An exception to this situation, the Barclay Psychiatric Hospital citation issued by OSHA required adjustments to staffing based on patient acuity.

Studying the impact of staffing on workplace violence is also difficult for researchers because employers are unlikely to provide access to at-risk workers if the research outcome may include recommendations for increased staffing.

Characteristics of an effective joint program 

Labor unions have used a variety of strategies to reduce workplace violence, including forming joint labor/management programs, filing complaints with enforcement agencies, negotiating contract language, and working for passage of state and federal legislation. In this author’s experience, cooperative programs in which management voluntarily “does the right thing” are usually the most effective. These programs are positive and usually experience less management resistance than actions imposed on management as the result of a grievance or citation from an outside agency. However, cooperative programs require strong management commitment, good labor relations, meaningful worker participation, and healthy organizational culture. These positive factors are rare in today’s workplaces, which are more often hierarchical and focused on the bottom line. Characteristics of an effective joint program include:

Rarely will a single quick fix exist that has a dramatic impact on reducing workplace violence. More typically, this involves a continuous process of identifying factors and making corrections. For example, in projects conducted by state worker unions and the New York State Office of Mental Health,6 improvements included adding telephones to the day rooms to reduce conflict over access to phones, use of mortise locks (automatic door locks) in place of skeleton-key locks, and instituting state-of-the art personal alarm systems.

Grievances, OSHA complaints, legislation 

When unions cannot gain cooperation from management in dealing with workplace violence, traditional adversarial tactics may be employed, including filing grievances, OSHA complaints, mass petitions, press actions, and working for new legislation.

Education and training initiatives 

Labor unions, employers, federal OSHA, states, and many professional societies have participated in training and education initiatives on workplace violence. The United Food and Commercial Workers, for example, has been successful in using OSHA grant funds to train workers and employers in commercial food establishments, particularly late-night and 24-hour workplaces. These programs seek to provide the tools necessary to implement joint labor/management programs that identify risk factors for violence and to put in place appropriate hazard controls.

Employers have also used training as a primary strategy to prevent workplace violence. For example, in public-sector institutions, numerous training programs have been designed to train staff to use nonverbal, verbal, and physical intervention skills to prevent and manage the violent behavior of people in their care. Many of these programs focus on the individual skill of the worker being trained and do not take the holistic approach outlined in the federal OSHA guidelines.

Research needs 

Intervention research that tests OSHA guidelines by instituting comprehensive violence prevention programs in high-hazard industries is a top priority to labor. These projects have the benefit of implementing violence prevention programs in affected work environments. Further, these programs can serve as a model that can be expanded to other sites covered by a particular collective-bargaining agreement or industry. Intervention research also helps lay the groundwork for promulgating effective OSHA standards. A key element in rule making is to demonstrate that proposed rules are technically and economically feasible. The cooperation that industry, labor, and researchers develop in the course of such projects should help improve relationships among these parties. For example, the University of Maryland, the New York State Public Employees Federation, CSEA, and the New York State Office of Mental Health were recently awarded a 3-year NIOSH National Occupational Research Agenda (NORA) grant to conduct an intervention study. Under the direction of an experienced researcher, this project will comprehensively study the implementation of OSHA’s violence prevention guidelines in four state mental health facilities. The study features statewide and facility-based project advisory committees and includes an economic analysis of the costs and benefits of the intervention.

In the context of research, a participative action research (PAR) model is preferable.31 A PAR features a process in which researchers work closely with representatives of the groups they study to inform the study’s design, methods, and implementation. When experts dominate research, they may become detached from the organizations they study. Are they studying the correct questions and are they missing nuances because they have ignored key stakeholders? Do recommendations reflect the real-world perspective of front-line workers, union representatives, supervisors, middle managers, and administrators? By involving the union and other groups in decision making, the flexible PAR process results in a better study design and buy in by labor and management, increasing the likelihood that the recommendations will be thoroughly implemented. If workers see researchers as relating solely to management, workers may mistrust the researchers’ findings and recommendations.

In today’s environment of sparse resources, intervention research projects may be attractive to labor/management groups that struggle to find time, personnel, and funding to implement programs. Researchers should keep in mind that unions and employers will be more likely to welcome research that can help them solve problems in the course of studying important phenomena.

Labor unions can provide a gateway to accessing many worker populations at high risk for workplace violence. In the public sector, this category includes schoolteachers; health care workers in state and county institutions for mentally ill, mentally retarded and developmentally disabled; nursing home workers; criminal justice workers; and social service workers. In the private sector, at-risk retail food workers have a rich history of union organization and are a prime population to participate in workplace violence–intervention research. Researchers should be aware of important differences from traditional research such as:

Detailed recommendations 

Conclusion 

Researching effective interventions to curb the devastating effects of workplace violence should become a top priority in the occupational safety and health community. Employers, unions, health and safety specialists, and researchers, especially those involved in high-risk industries, should explore every opportunity to collaborate in this pressing area of work. The lives and well-being of workers, families, and the businesses they work for depend on it.