The Changing Organization of Work and the Safety and Health of Working People

Knowledge Gaps and Research Directions

Chapter 4: Intervention Research

• Gaps in Research on Organization of Work Interventions to Protect Safety and Health
• Intervention Research Needs
  • Effects of Protective Practices and Policies
  • Improvement in Intervention Research Methods
  • Implementing Interventions

Gaps in Research on Organization of Work Interventions to Protect Safety and Health

The scientific literature provides relatively few examples of occupational safety and health interventions that feature the reorganization of work. One exception is a fairly small body of research on reducing job stress that looks at effects of interventions such as work rescheduling, workload reduction, role clarification, and the redesign of jobs to improve worker decision-making and autonomy (see Parkes and Sparkes [1998] for a review of this literature). This research base is substantially enlarged if the broad literature is considered on sociotechnical design that examines effects of organizational interventions on related outcomes such as employee satisfaction and morale [Ilgen 1990; Locke and Schweiger 1979; Schneider 1985]. Additionally, studies of occupational safety programs are found in the safety literature [Cohen and Colligan 1998; Hugentobler et al. 1990; OSHA 1998] although classifying many of these programs as organization of work interventions may be inappropriate.

Review of this intervention research literature leads to several broad conclusions. First, as noted, the body of literature on interventions to change aspects of job design or organizational practices to reduce exposures to job hazards is small. In the job stress arena, this research base is much smaller than the body of research on individual-level intervention strategies, such as stress management and health promotion, that seek to improve the capacity of workers to withstand demanding or hazardous job situations [Murphy 1996].

Second, the extent to which many of these organization of work interventions improved worker safety and health is questionable. Small, inconsistent, and short-term effects of interventions are commonly reported, although these problems are more often found in the health literature than in the safety intervention literature. In their recent review of job stress intervention studies, Parkes and Sparkes [1998] concluded that “. . . the studies do not present a convincing picture of the value of organizational interventions designed to reduce work stress . . . and tend to be difficult to interpret, causally ambiguous, inconsistent, based on small samples, and/or statistically nonsignificant”. These conclusions were particularly true of participatory action research interventions; of the five such studies reviewed, not one demonstrated strong and favorable results.

Third, the organization of work intervention literature is beset by various methodological problems, especially the absence of strong study designs involving randomized trials, making evaluation and attribution of outcomes difficult. As discussed in a recent commentary on research on prevention of occupational injuries [Rosenstock and Thacker 2000], such methodological shortcomings are pervasive in the occupational safety and health intervention literature.

On the other hand, strong designs are sometimes not feasible in intervention research because of practical, ethical, legal, or other constraints. For example, organizations are changing so quickly in response to technological innovations and market conditions, it is often difficult to achieve the degree of stability needed to maintain control groups essential to experimental evaluation of interventions. At a more theoretical level, it has also been argued that organizations cannot be readily changed or improved by attempts to systematically manipulate their individual parts, and thus conventional scientific approaches may not be faithful to the true dynamics of the organizational change process [Colarelli 1998]. These circumstances represent substantial obstacles to rigorous study designs in intervention research and have led to the adoption of quasi-experimental methods in some cases and to the use of action research methods in others.

Finally, in addition to gaps in intervention effectiveness research, little is known about implementing interventions within organizations. Evidence suggests that benchmarking and other practices that are not necessarily evidence-based play an important role in adopting and diffusing organizational interventions (in contrast to the public health model, which proceeds methodically from problem identification to intervention and evaluation). However, these practices are not well understood. Absence of better information about the decision processes underlying organizational interventions thus becomes an important impediment to promoting safety and health interventions for workers.

Intervention Research Needs

To close these gaps in research on organization of work interventions, advances are needed along several fronts: (1) a substantial increase in research is needed on organizational practices and policies that may serve to protect worker safety and health; (2) methodological advances in intervention research are needed in support of these studies; and (3) a closer examination is needed of factors that influence the implementation of interventions within organizations.

Effects of Protective Practices and Policies

Three categories of intervention research needs are most evident. First, intervention research is needed that develops remedies to hazards identified in health effects studies of the changing organization of work. Naturally, this research should target those organizational changes that create the highest apparent risk for stress, illness, and injury. In this regard, studies of the safety and health benefits of interventions targeting practices such as organizational restructuring/downsizing, long hours of work, and work intensification could be recommended.

Second, a need exists to more aggressively investigate the effects of organizational interventions already occurring in the workplace (i.e., natural experiments) to protect the well-being of workers. For example, work-life programs and family friendly policies such as flexitime and flexiplace are becoming increasingly common in the modern workplace, presenting opportunities for research to investigate effects of these types of interventions on worker safety and health. This type of research may be more properly called program evaluation.

Finally, increased research attention needs to be given to the effects of legislation and public policy that influences the organization of work. One example of such legislation is the Family Medical Leave Act (FMLA), which ensures the right for leave time for critical family medical reasons. It can be anticipated that the FMLA would reduce stress associated with added workload burdens, job insecurity, or work-family conflict that may otherwise result from the need to provide family medical care. However, empirical study of the FMLA or other social policy impacting the workplace is scarce.

All these types of intervention studies should include effectiveness measures for a wide range of cost factors such as illness, injury and disability rates, health care utilization, absenteeism and lost time, etc., in addition to measures of health and well-being that are more commonly collected in research on the organization of work (e.g., selfreports of health status, job satisfaction). Such measures will help to better understand the intervention impact on injury and disease burden sustained both by workers and the organization.

Along this line, a new concept of organizational effectiveness (sometimes dubbed “organizational health”) links organizational practices that protect worker safety and health with high levels of organizational functioning—a variant on the theme “safety pays”. Safety and health studies of organizational interventions that examine a broad range of organizational outcomes may identify practices that are conducive to both improved worker safety and health and improved organizational effectiveness, thereby building a stronger case for worker protection through job redesign.

As a practical matter, advances in organization of work intervention research would benefit from improved resources and training. A need exists for developing and compiling information about the science of intervention research, and for increased exposure of researchers to this information in their graduate training to improve their capacity for intervention research. In this regard, it is notable that the topic of intervention effectiveness has been recognized as a research priority under NORA, and a team of specialists from NIOSH and outside interest groups have undertaken an evaluation of knowledge on this topic similar to the present exercise for organization of work [Goldenhar et al. 2001].

Improvements in Intervention Research Methods

Advances in knowledge of organizational interventions to protect worker safety and health will require improvements in intervention research practices. There are two major issues here. First, a need exists for improved guidance, and possibly a new paradigm, for designing and conducting research on organizational interventions. Stated in the most general terms, approaches need to be articulated that build on the strengths and minimize the limitations of various contrasting methods (e.g., case studies and action research versus experimental designs, qualitative versus quantitative methods) and allow for expeditious conduct of interventions in an organizational context. Responding to this need, Zwerling et al. [1997] suggested that a sensible and economical approach to conducting intervention effectiveness studies begins with qualitative methods and quasi-experimental designs to explore the feasibility of interventions, holding randomized controlled trials in reserve for testing and validating the most promising approaches.

Second, intervention research needs to be more theory driven to learn why and under what circumstances organization of work interventions succeed. In their recent report, the NORA intervention effectiveness team noted that too often investigators do not adequately describe the intervention or address the issue of why or how an intervention is expected to bring about improvement in the safety and health of workers [Goldenhar et al. 2001]. Well designed, theory-driven intervention research increases the likelihood that an intervention will ultimately be effective because it leads to a better understanding of how the intervention works and allows for generalization and tailoring of the intervention to multiple situations.

Until recently, theory-driven models for occupational safety and health interventions were not common in the published literature. However, Cohen et al. [1997] developed a model for reducing work-related musculoskeletal disorders that prescribed a seven-step process beginning with problem identification and leading ultimately to the design of new work practices (intervention). Melhorn et al. [1999] used this model to establish an occupational intervention program for musculoskeletal disorders and reported substantial savings in worker compensation costs. Israel et al. [1996] presented a conceptual framework for interventions to reduce stress at work that was based on a comprehensive model of stress and health. Such models provide guidance for each step in the intervention process (design, implementation, and evaluation). As a side note, both of the above-mentioned models emphasize the need for management commitment and employee involvement in the intervention design process.

Implementing Interventions

More information is needed about critical factors and conditions that motivate organization of work interventions to protect worker safety and health. Experience suggests that decisions for organizational change are often driven by industry opinion leaders, by authority figures in organizations who are convinced about the efficacy of new organizational practices, by best practices, or by industry norms (benchmarking). However, these decision processes are not well understood or appreciated in the scientific community.

Ethnographic study of organizations to better understand the processes that govern intervention decisions could lead to development of products to help motivate and support interventions. Examples of these products might include (1) casebooks on successful interventions to help organizations select, guide, and evaluate interventions; (2) design and promulgation of best practices based on accumulated findings from intervention research; and (3) forums on organization of work and health that are keyed to practitioners (in contrast to researchers) and highlight information about interventions and intervention effectiveness. Study of these decision processes could also lead to improved designs for intervention research (e.g., help to identify organizationally relevant measures of intervention effectiveness).