Research impact is often assessed through statistical analysis of scientific publications citing the work or through qualitative assessments by subject-matter experts. However, such assessments seldom estimate the economic benefit associated with particular research investments in terms of lives saved, injuries or illnesses averted, or increases in worker productivity maintained. The National Institute for Occupational Safety and Health (NIOSH) asked the RAND Corporation to develop and illustrate an approach for estimating the economic benefit of NIOSH research, using three case studies. The cases provide concrete illustrations of the ways in which NIOSH research might have an impact on worker health and safety practices and outcomes, as well as some initial estimates of the economic benefit associated with those impacts.
We selected the case studies to illustrate variation in types of NIOSH research and in intended users. The first case study examines research to develop, test, and support implementation of engineering control measures to limit exposure to silica among road construction workers and offers an example of NIOSH's intervention and surveillance research and provision of technical assistance. The second case study involves two NIOSH studies that strengthened the evidence base about the linkage between firefighting activities and increased risk of certain cancers among firefighters and provides an example of etiological and exposure surveillance research, coupled with an intervention study. The third case study involves a NIOSH evaluation of the effectiveness of Ohio's Safety Intervention Grant Program and implementation of safety-oriented engineering controls by employers and illustrates intervention research targeting government organizations. The research in the first and second case studies led to the development of control technologies, and all three case studies involved dissemination and stakeholder engagement efforts that promoted the adoption of risk-reducing technologies and practices.
Assessing the economic benefit of such research requires assigning a dollar value to prevented injuries, illnesses, or deaths using risk-reduction measures derived from the research; determining whether such risk-reduction measures might have occurred without the research in question; and determining whether a particular entity (e.g., NIOSH) made a significant contribution to any resulting benefits. Doing so is fraught with difficulties, including the fact that the benefits of research can occur many years in the future, the absence of natural market price mechanisms for many outcomes, and the difficulty of assessing the contributions of research by any one organization, such as NIOSH, to any observed benefits. The case studies walk the reader through a transparent set of logical steps, marshaling quantitative estimates where possible, and providing transparent discussion about our assumptions where such evidence is not available.
The study employed two common approaches to estimating the economic benefit of avoided injuries, illnesses, and fatalities. The first involves estimating associated medical costs and productivity losses, which is often useful in addressing questions related to budgeting for medical care and other costs. However, where there were gaps in available cost data, we employed a second approach, which expresses benefit in terms of value of a statistical life (VSL) and attempts to take a broader societal perspective than the first approach and value all costs to society, whether “on budget” or not. Given this approach's broader scope, VSL estimates tend to be significantly larger than medical costs. With that in mind, the key findings for each case study are as follows:
There are limitations to this analysis. First, given limitations in available data, we found it necessary to make some important assumptions in order to derive estimates of the economic benefit. Where we did this, we explained the logic behind the assumption and, in many cases, conducted sensitivity analyses to clearly show the reader how different assumptions might affect the estimated benefit. Second, the reader should bear in mind that differences in estimated benefit reflect differences not just in effectiveness but also in maturity of the field and the state of prior knowledge. Adjusting our estimates based on such differences among our three case studies was beyond the scope of this study. Third, providing estimates of costs, which are necessary for determining whether benefits are “large enough,” was beyond the scope of this analysis.2 A final caution is that, given the project timeline, we selected the three case studies in part because of data availability. Hence, although the findings might help stimulate discussion, they do not provide a definitive assessment of NIOSH's overall impact.
In spite of these limitations, the value of this analysis lies in illustrating some specific ways in which NIOSH research can produce economic benefits, providing some sense of the likely magnitude of these benefits in dollar terms, and providing an analytical framework on which others can build in future work.
In the future, NIOSH should consider conducting additional case studies to explore other types of research and intended audiences and that account for the costs of producing and implementing research. In addition, it should consider examining cases in which the linkages between its research and safety and health improvements are less clear because there can be important lessons from cases of unrealized impact. Finally, NIOSH should consider ways in which it might start to fill some of the gaps in data and analysis encountered while conducting this economic analysis.