Moving forward with research on the prevention of violence against workers

Introduction 

The papers that came from and the discussions that took place within the Workplace Violence Intervention Research Workshop, sponsored by the University of Iowa Injury Prevention Research Center in April 2000, demonstrated the mismatch between the magnitude of the problem and current knowledge about prevention. With more than 700 workplace homicides1 and nearly 2 million nonfatal assaults in 1998,2 workplace violence poses a public health problem that employers also need to embrace as a significant concern. Yet, efforts to ameliorate the problem have been limited in scope and have received little support for epidemiologic studies or research testing interventions.

Future progress will rely on addressing several categories of issues identified in the meeting, including:

Definition and measurement of workplace violence 

Violence against workers occurs within diverse contexts. These encompass such disparate situations as robbery-related homicides in retail businesses; homicides of taxi drivers; attacks on teachers by students; assaults on workers by patients being treated in psychiatric facilities; assaults on law enforcement officers; attacks on health care or social services providers working with troubled clients or in high crime neighborhoods, often under highly volatile conditions. Violence against workers may also result from disputes among coworkers or from partner violence that spills over into the work environment. Understanding the problem to develop appropriate interventions requires a common vocabulary to ensure clear and consistent measurement of exposures and outcomes across settings and studies. Developing specific categories of violence against workers3, 4 is an important contribution to the field, but does not resolve all the issues associated with definition and measurement. For example, efforts must be made to standardize definitions of assaults across settings and data systems and to develop a clear vocabulary for describing intervention modalities.

Improving our understanding of the circumstances in which violence against workers occurs will help us develop intervention strategies. Ideally, the epidemiology of violence against workers and also increased understanding of intervention research in other content areas will inform future research. This research must incorporate theories and frameworks from the social sciences and health behavior fields; from environmental sciences, including industrial hygiene; and from business, including expertise about management, organizational behavior, and industrial relations.

Strategies for prevention 

Strategies to prevent violence against workers must be diverse. However, common principles can undergird intervention development. Haddon5, 6, 7 developed principles of injury prevention that, although applicable here, have not been used widely. For example, the Haddon matrix applies a socio-ecologic framework to develop interventions directed at individuals or at the vehicles by which injury-producing energy is transmitted (e.g., modifying products like guns or equipment). Likewise, the Haddon matrix focuses on the broader social and physical environments in which injury occurs and that contribute to injury. As such, it delineates a number of potential targets for intervention. In addition, the matrix provides a format to consider interventions targeted at different stages of violence and injury, including strategies directed at preventing assaults, reducing injuries associated with assaults, alleviating the long-term consequences of assault or injury, and preventing future incidents (see Table 1).

Table 1. Haddon matrix—examples of potential strategies to prevent violence against workers

	Phases	Factors
		Host (victim)	Agent/vehicle/vector (weapons & assailant)	Physical environment (structures & facilities)	Social environment (norms, policies & procedures)
	Pre-event (pre-assault)	Train workers to identify potentially violent clients or customers Train managers in conflict resolution and proper dismissal strategies	Make weapons less easily concealed (weapon) Provide careful oversight of potentially volatile employees (assailant) Educate patients/clients in anger management (assailant)	Modify structures to decrease ease of access by unauthorized persons (e.g., fired workers or violent partners) Install metal detectors Install bulletproof shields between workers and customers	Reduce access to weapons Prohibit solo workers in high-risk establishments
	Event (assault)	Train workers in methods of signaling for help during robberies and assaults Train employees to use self-protection measures when confronted with violent client	Reduce lethality of weapons (e.g., fewer firing rounds, less lethal bullets)	Install and maintain easy-to-operate alarm systems Reduce isolation of work spaces Ensure that workers have escape route	Ensure adequate security backup for threatened workers Develop plan for responding to threats
	Post-event (post-assault)	Train workers in first aid Provide workers with crisis-intervention counseling after assault events	Reassign workers after coworker violence Improve ability to trace firearms and apprehend suspects	Ensure access to the work site by emergency vehicles Install cameras to facilitate identification and apprehension of assailants	Provide adequate insurance plan to workers for acute and long-term medical care and counseling services

Haddon’s other framework defines ten countermeasures for reducing injury,6 as listed in Table 2. This formulation, similar to concepts in industrial hygiene, can also direct the development of interventions in ways that may not be immediately apparent. For example, one method of reducing robbery-related homicides in convenience stores might be to eliminate the late-night operation of these businesses. This, like all interventions, must be carefully considered in terms of the political and social tradeoffs required. However, such breadth of ideas requires inclusion in the debate.

Table 2. Examples of the Haddon countermeasures applied to preventing violence against workers

	Haddon countermeasures6	Examples applied to violence against workers
	1. Prevent creation of the hazard	• Eliminate late-night retail operations
	2. Reduce the amount of hazard brought into being	• Reduce the number of violent patients/clients under the care of a given worker
	3. Prevent the release of the hazard	• Isolate or restrain violent patients or prisoners
	4. Modify the rate of release of the hazard from its source	• Eliminate automatic weapons
	5. Separate the hazard from those to be protected by time and space	• Assign workers having conflicts to different shifts • Move worker being threatened by a partner to a different work site
	6. Separate the hazard from those to be protected by a physical barrier	• Install bullet proof shields in taxis and police vehicles, and between cashiers and customers in retail settings
	7. Modify relevant basic qualities of the hazard	• Manufacture fewer lethal weapons • Provide counseling to those with post-traumatic stress syndrome who may be prone to violence
	8, Make those to be protected resistant	• Train workers to handle assaultive patients or customers • Provide protective equipment to workers
	9. Begin to counter damage done by the hazard	• Provide counseling and health care for workers after assault
	10. Stabilize, repair, and rehabilitate the object of the damage	• Provide medical care and rehabilitation services to injured workers

Runyan8 has proposed a strategy for weighing the options derived from Haddon’s frameworks. The strategy requires comparing and assessing the various attributes of possible interventions relative to one another. Without a framework for generating intervention ideas and without data for assessing the attributes of various options, progress will be limited. Developing the information base to permit the kind of assessment described in Runyan8 will require additional data collection and careful analyses.

After designing interventions that incorporate principles of behavioral and physical sciences, it is critical to develop appropriate evaluation strategies. All too often evaluations are too narrowly defined, poorly conceptualized, and poorly executed. A recent special issue of this journal included reviews of occupational injury–intervention literature in 12 topical areas.9 It is troublesome that few high-quality evaluations of interventions have been conducted in any area of occupational injury prevention, including violence.7, 10

Often the impetus for evaluation involves guiding resource allocations, asking only, “Does it work?” or “Is it cost-effective?” This limited focus on outcome, and possibly cost, fails to consider intermediate impacts or program process. Such evaluations are often inadequate and will yield limited results because the effects measured (e.g., reductions in homicides or in nonfatal injuries) require much more time to become evident than the funded evaluation period allows. Consequently, the evaluation fails to demonstrate positive outcomes, and potentially useful interventions may be abandoned.

Impact measurements are frequently more realistic in a limited time frame and should not be discounted. Such evaluations of intermediate outcomes are valuable, but still require clear definitions and measurement. For example, impact evaluation would assess more intermediate outcomes, such as adopting a policy or implementing a safety strategy, rather than relying on changes in morbidity or mortality. The value of impact evaluations depends to some degree on the strength of relationship between the impact measured and the intended outcome.

Rarely have evaluations in this topic area included careful assessment of process. This type of evaluation helps one understand why an intervention may have failed, or what elements of the intervention are more difficult to transfer across settings, accounting for the failure in one setting of an intervention successful in a different environment. Likewise, process evaluation delineates what actually occurred during the intervention and can frequently help to identify modifications in procedures that may have affected the success or failure of the intervention. Examples include use of a trainer with different experience or skills than originally planned, problems in attendance at an educational program or a briefing critical to explaining a policy strategy, or language barriers between the interventionist and the intended recipients of the messages.

Future development of successful strategies for preventing violence against workers will depend on the ability of multidisciplinary scholars and practitioners to work together to consider various perspectives in conceptualizing problems, solutions, and evaluations. This means carefully incorporating the interests of labor and management, and it means that epidemiologists, industrial hygienists, and behavior-change specialists must search for commonalties in perspectives and figure out where their approaches converge and where solutions will be strengthened by distinct, but complementary approaches. Likewise, the differences in perspective among professionals engaged in law enforcement, public health, media, industrial psychology, and public policy must be identified in complementary terms and viewed as assets, rather than as barriers to communication, collaboration, and progress.

Recommendations 

The next steps in addressing violence against workers must be supported adequately, not only through conceptual work that merges perspectives across disciplines but also by ensuring adequate resources to study violence against workers, developing appropriate intervention responses, and evaluating them fully. Recommendations follow:

Conclusion 

This meeting pointed to the need for substantially more research funding and attention to the broad problem of preventing violence against workers. The discussion and specific recommendations from the meeting begin to chart the course for this research. To successfully understand the problem and to derive appropriate and replicable solutions, researchers must factor into their approach a variety of concerns, as outlined here. Investment in this research promises not only to reduce workplace violence but also to improve efforts in preventing violence outside of work environments and to further the safety of workers associated with other hazards.