MIKE JAYJOCK, PhD, CIH, FAIHA, is a consultant at Jayjock Associates LLC based in Langhorne, Pennsylvania.
NEIL C. HAWKINS, ScD, is president of the Erb Family Foundation and former vice president of EHS and chief sustainability officer for Dow Chemical. Acknowledgements: The authors thank Kathy Fehrenbacher, Sheldon Lande, Jerry Lynch, Pertti Hakkinen, Mark Nicas, and others who were inspirations and visionaries. Send feedback to The Synergist.
Exposure Modeling: The Next Generation
Over 30 years ago, when risk assessment and risk management (RA/RM) was becoming the key decision paradigm for the evaluation and control of agents hazardous to human health, we began to focus on developing the science of exposure assessment and exposure modeling. This was unusual because most exposure assessment and risk assessment were environmental, and not part of the IH toolbox. Within the RA/RM paradigm, our specialty was exposure modeling, a topic about which we have, together and separately, published seven papers since 1988.
During those early days, exposure sampling and monitoring dominated everything in the field, but there was nascent interest in exposure modeling. Rudimentary exposure modeling had been developed in the early 1980s by contractors used by EPA for exposure assessment of new chemical substances under the Toxic Substances Control Act. But OSHA and NIOSH, in our judgment, never showed much interest in modeling. NIOSH focused on the use of historical data in retrospective exposure assessments in epidemiology studies. The use of exposure models in a workplace setting for contemporaneous decisions was essentially unheard of. Few models existed, and those that did were not trusted for safety and health decisions of ongoing, actual exposures.
In addition to EPA contractors, some academic researchers were also exploring the use of exposure models. More importantly, in the private sector we saw the practical need for models that would organize and optimize our monitoring and RA/RM efforts. Many in industry started to recognize the strengths and weaknesses of monitoring and RA/RM; we clearly understood that we would never be able to measure and sample everything, everywhere.
MODELING VS. MONITORING We were also beginning to appreciate that modeling could provide a more scientific basis for exposure assessment and the practice of industrial hygiene. Models formalize and document the formation of hypotheses regarding human exposure. Except on a superficial level, monitoring cannot serve this function. Modeling also defines and discloses the quantitative relationship between exposure and its independent physical drivers. Monitoring, in turn, contributes to the evaluation, validity, and further development of hypotheses. Modeling can—and perhaps should—provide the context for sampling results.
During this early era, both modeling and monitoring were deterministic, with little recognition of or allowance for the uncertainty in estimates. Exposure data was obtained via the collection of a few samples largely driven by the need for regulatory compliance and to compare measured exposures to exposure standards. There was little focus on uncertainty in sampling-oriented exposure assessments to drive research needs or to facilitate confident risk management decisions.
In 1995, we published a paper in Applied Occupational and Environmental Hygiene that urged quantitative evaluation of the drivers of exposure to facilitate model development. We also continued to push for validation studies for models, assessment of uncertainty, best practices for exposure assessment, and complete transparency—topics that we had explored in papers published by the AIHA Journal in 1991 and 1993. We always saw the need for both sampling and validated modeling of exposures with a risk management framework and assessment of uncertainty.
Models formalize and document the formation of hypotheses regarding human exposure. Except on a superficial level, monitoring cannot serve this function.
VISION VS. REALITY Clearly, some progress has been made in both modeling and monitoring over the last 30 years. Occupational exposure assessment made a great leap forward with the publication of AIHA’s first monograph in 1991. The third edition of The Occupational Environment—Its Evaluation and Control in 2011 helped promote the use of statistical analysis and tools, such as AIHA’s IHSTAT spreadsheet, that provide for a measure of uncertainty in monitoring data. The universe of first-principle exposure models was described in the second edition of AIHA’s Mathematical Models for Estimating Occupational Exposure to Chemicals, published in 2009. Free software tools to ease the computational tasks associated with the models—the IHSTAT, IHSkinPerm, and IHMOD spreadsheets—are available from AIHA's website. Some, albeit very few, validation studies of exposure assessment models have been published. Also, model uncertainty analysis has been properly demonstrated and can be achieved through use of practical tools such as software for Monte Carlo analysis of the input parameters for exposure assessment models.
But our vision over 30 years ago was that today’s industrial hygienists could take a validated model off the shelf for a typical exposure scenario and produce, with confidence, an exposure assessment with transparency and a real description of uncertainty for a risk manager making a corporate or regulatory decision. This vision is clearly not today’s reality.
So, what is preventing the use of exposure models? Our answer to this question is simple: we do not have the datasets needed to feed and validate the models. The cost of obtaining these datasets is not trivial, but they will be generally and eternally useful to the proper development of the science of exposure assessment. The feasibility and value of supporting the development of these datasets was demonstrated in “Evaluating Well-Mixed Room and Near-Field-Far-Field Model Performance Under Highly Controlled Conditions,” a paper that appeared in the Journal of Occupational and Environmental Hygiene in 2017.
To advance the use of modeling, we make the following recommendations to the exposure assessment community: •Recognize that occupational exposure banding and the use of empirical exposure models such as Stoffenmanager and the Advanced REACH Tool are interim, stop-gap approaches to exposure assessment; they are not the final answer to confident and cost-effective exposure assessment in the service of human RA/RM.
•Prioritize funding the development of airborne contaminant emission and depuration databases and subsequent submodels as a worldwide public works project.
•Review the comprehensive research outline for these factors, which was crafted by global modeling experts in a two-day European Union workshop over 16 years ago (PDF).
•Use the EU document as a starting template for a funded research program for emissions.
We raise these issues because the practice of exposure assessment has not advanced as far as necessary, and a new generation needs to take the torch. But we are optimistic because the power and utility of exposure modeling have been clearly illustrated during the aerosol exposure challenges related to COVID-19. The world has seen how modeling can help inform risk management decisions. Let’s capitalize on this momentum and create the right environment for occupational exposure modeling for the next generation.
AIHA: “AIHA Risk Assessment Tools.”
AIHA: A Strategy for Occupational Exposure Assessment (1991).
AIHA: Mathematical Models for Estimating Occupational Exposure to Chemicals, 2nd ed. (2009).
AIHA: The Occupational Environment—Its Evaluation and Control, 3rd ed. (2011).
AIHA: “Uncertainty” in Mathematical Models for Estimating Occupational Exposure to Chemicals, 2nd ed., chapter 10 (2009).
AIHA Journal: “Assessment of Inhalation Exposure Potential from Vapors in the Workplace” (August 1988).
AIHA Journal: “Monte Carlo Uncertainty Analysis of a Diffusion Model for the Assessment of Halogen Gas Exposure During Dosing of Brominators” (June 1997).
AIHA Journal: “A Proposal for Improving the Role of Exposure Modeling in Risk Assessment” (December 1993).
AIHA Journal: “A Rationale and Framework for Establishing the Quality of Human Exposure Assessments” (January 1991).
AIHA Journal: “Uncertainty Analysis in the Evaluation of Exposure” (May 1997).
Annals of Occupational Hygiene: “Advanced REACH Tool (ART): Overview of Version 1.0 and Research Needs” (November 2011).
Annals of Occupational Hygiene: “Stoffenmanager Exposure Model: Development of a Quantitative Algorithm” (August 2008).
Applied Occupational and Environmental Hygiene: “Exposure Database Improvements for Indoor Air Model Validation” (April 1995).
EPA: “A Manual for the Preparation of Engineering Assessments” (February 1991).
Journal of Occupational and Environmental Hygiene: “Evaluating Well-Mixed Room and Near-Field-Far-Field Model Performance Under Highly Controlled Conditions” (June 2017).