The key is to select a model that is simple enough to use but can provide a relevant answer to support an appropriate decision.
AIHA’s launch of the Product Stewardship Society in 2012 was in part an acknowledgement of the industrial hygienist’s core skills relating to exposure and risk assessments (EA/RA), not only for primary production workers but for a broader range of potentially exposed populations, including downstream industrial and commercial workers, consumers, and the general population. These exposures stem from uses of products and subsequent releases to air, water, soil, wildlife, and food production. This article introduces some of the tools industrial hygienists can bring to EA/RA for downstream product use.

Exposure assessments for both manufacturing and product use begin by looking at the tasks workers or users perform that put them in potential contact with the hazard. This evaluation may require an understanding of the physical and chemical properties of the hazard; the nature and severity of its toxicity; and the frequency, duration, and intensity of the contact, possibly by all relevant exposure routes, as well as the environmental conditions (temperature, ventilation, control technologies) during use. The exposure evaluation could include professional judgment, exposure/control banding, screening measurements, more detailed measurements, or mathematical modeling. The results might be compared qualitatively or quantitatively to an occupational exposure limit.
For occupational exposures, our goal is typically to determine if they are acceptable as a ceiling limit, a short-term exposure limit (STEL), or a time-weighted average (TWA) for an adult worker throughout a full working career. But exposures for consumers and the general population aren’t limited to a working day. They are assessed over various times—a short-term application, a 24-hour period, or an entire lifetime.
Exposure assessments for manufacturing environments differ in several ways from those for downstream users of consumer products. At the start, the relevant populations are characterized through scenarios describing the product’s properties; the quantity, frequency, and duration of its use; and the environmental conditions. In occupational settings, we consider many types of hazardous endpoints, including CMT (carcinogen, mutagen, and teratogen) exposures and risks. These may be of even greater concern in general population and consumer applications. Efficacy of exposure controls in downstream and consumer uses can be quite different than in primary production operations. Controls against inhalation, ingestion, and dermal uptake are often difficult to ensure for consumers because we can’t assume they’ve received training and education on the hazards or that they’ll use the proper PPE.
Concerns about sensitive sub-populations are more extensive with consumers. Sub-populations include adults in childbearing years and children with in-residence exposure for nearly 24 hours a day. Children have much different activity patterns than workers or general population adults; for example, infants crawl on floors and carpets and exhibit frequent hand-to-mouth behavior. Oral exposure is less frequently an occupational exposure concern.
For assessments of consumers and the general population, the exposure assessor must understand which population segments are potentially most at risk and their “exposure factors” such as body weight, breathing rate, hand-to-mouth behavior, and dermal contact/uptake, as well as the composition, frequency, duration, and intensity of the exposure. The allowable exposure guidelines, which include but are not limited to an Acceptable Daily Dose (ADD) and Allowable Daily Intake (ADI), must then reflect the possibly unique characteristics of the at-risk populations.
Editor’s note:
This is the third article in a four-part series on mathematical modeling in exposure assessment, following “
Patterns of Exposure
” by Chris Keil (January 2017) and “
Easy Modeling
” by Paul Hewett (April 2017). The final article, to be published in December, will discuss models for emergency response planning.
Figure 1.
Inputs and results from IH Mod for a well-mixed room model for a hypothetical exposure.
Figure 2.
Results from the near-field/far-field model are slightly higher than for the well-mixed room model.

Tap on the figures to open larger versions in your browser.
These estimates have several uncertainties. For example, the entirety of a semi-volatile substance in a formulation like caulking compound doesn’t instantaneously emit when applied. Whatever amount of X is on the fresh surface layer of the caulk may evaporate quickly, but the substance in the lower microlayers must first diffuse to the surface before evaporating. A moment or so after application, the curing film on the caulk surface starts limiting evaporation. These processes are quite difficult to model accurately, and some experiments designed to measure the loss of the substance over time may be invaluable in refining the exposure and risk estimates. However, the likely slower “real world” rate means our “all emitted as applied” estimate is biased high in the short term. It also means there is a prolonged exponential decay in the emissions from the applied caulk.
Often, the goal of product exposure assessment is a dose that can be compared to an Acceptable Daily Dose (ADD) or a Lifetime Average Daily Dose (LADD). Let’s assume that Substance X has potential for developmental toxicity, with an ADD of 1 mg/kg/day. A young woman applies the caulk with some dermal contact as well as inhalation exposure. Looking to EPA’s
Exposure Factors Handbook
, we find a median moderate activity breathing rate of 2.2E-2 m3/min for women aged 21 to 31 (Table 6-19) and a median body weight of 68 kg for women between 21 and 29 (Table 8-5). Using the near-field TWA for the 45-minute exposure of 63 mg/m3 and assuming 100 percent absorption from inhalation, the inhalation dose is then 63 mg/m3 * (0.022 m3/min * 45 min) = 62 mg for a mg/kg dose of 0.92 mg/kg/event.
Dermal exposure is a concern, and a tiered approach works here, too. As a start, consider the dermal model in AIHA’s publication
A Strategy for Assessing and Managing Occupational Exposures
. For this model, we need to input the dermal surface area contacted, the quantity of substance retained on the skin, the percentage of chemical in the formulation applied, the number of contacts per day, and the absorption fraction. To simplify, we will assume dermal contact over sufficient skin surface to produce an applied mass of 5 percent of the product used, with about 4 grams to the skin (75 grams per application * 5 percent). We will also assume full uptake (100 percent absorption) of Substance X to arrive at 0.2 mg (4 grams * 5 percent) as a dermal dose estimate. Here, the inhalation dose during the application is the main source of exposure, but the combined dose is 1.1 mg/kg/event.
IH Mod can give a TWA over a whole day (1,440 minutes), which considers the purge of the space after application ceases, to allow estimation of a full 24-hour inhalation dose. With the 24-hour TWA of approximately 2 mg/m3 * (0.022 * 1,440), or 63 mg, and an additional 0.2 mg from dermal, the total dose for the application day is an estimated 0.93 mg/kg/day. Note the respiration rate may be lower on average if we include rest time, so this is possibly a high estimate.
Although there are numerous uncertainties in the models here, the predicted exposure is quite close to the guidance limit. It is arguably not very likely that more sophisticated exposure modeling would substantially reduce the predicted concentrations and dose and then change the decision that this product use scenario represents a use of probable concern.
Whether your exposure assessment concerns use of consumer products or the hazards of a traditional manufacturing environment, modeling is best done with help from others. If you have access to additional resources, always ask them to review your inputs and assumptions.
is principal at TWA8HR Occupational Hygiene Consulting, LLC. He can be reached via
A Strategy for Assessing and Managing Occupational Exposures
, 4th edition (2015).
Applied Occupational and Environmental Hygiene:
“A Tiered Approach to Deterministic Models for Indoor Air Exposures” (January 2000).
Consumer Product Information Database
. EPA:
Consolidated Human Activity Database
. EPA:
Exposure Factors Handbook
. EPA:
Methods for Assessing Exposure to Chemicals
, Volume 7: Methods for Assessing Consumer Exposure to Chemical Substances
(1987). European Centre for Ecotoxicology and Toxicology of Chemicals:
ECETOC Human Exposure Assessment Tools Database
. European Commission:
Scientific Tools and Databases
International Journal of Environmental Research and Public Health
: “Exploring Global Exposure Factors Resources for Use in Consumer Exposure Assessments” (July 2016). Soap and Detergent Association:
Exposure and Risk Screening Methods for Consumer Product Ingredients
, April 2005). Springer:
Residential Exposure Assessment
: A Sourcebook
The Synergist
: “
Making Decisions with Uncertain Data
(January 2015). U.S. Department of Health and Human Services:
Household Products Database
Modeling allows us to evaluate specific exposure scenarios, and can help us prevent risks before they’re introduced. One of our core decisions is to select the correct model. But there are many models. How do we choose?
The key is to select a model that is simple enough to use but can provide a relevant answer to support an appropriate decision. The model also needs to be able to answer the question at hand and be relevant to the exposure scenario and the substance of interest. For example, many models do not work well for particulates. Other models are specific to worker or consumer exposures. Some models are based on established tasks or activities that may not be relevant to the scenario you’re assessing.
Often it makes sense to start simple and go more complex if necessary. If a very simple and probably biased-high model gives a “no problem” answer, the process may end. But if more sophistication is required, a “tiered” approach, which some researchers have advocated for occupational modeling, can be applied to downstream product assessments, too. (See “Exploring Global Exposure Factors Resources for Use in Consumer Exposure Assessments” in the July 2016
International Journal of Environmental Research and Public Health
for more information about types of models and modeling resources.)
Another consideration is the type of exposure estimate needed. Does the scenario involve a flammable substance used in a confined space, where the goal is to keep exposures below a short-term percentage of the lower explosive limit? Is a ceiling limit, STEL, or TWA needed? Be aware of the model’s strengths and limitations with respect to these issues.
Let’s explore application of models to a hypothetical example of downstream formulation, commercial, and consumer uses of a product. For simplicity, we’ll use point estimates. As the EA progresses, conversion to data input ranges and probability distributions may be merited, and then stochastic probabilistic models (for example, Monte Carlo calculations) might provide useful results. (See “
Making Decisions with Uncertain Data
” in the January 2015
Consider a solvent (Substance X) manufactured in pure form by a specialty chemical manufacturer. Downstream, X is formulated into various products used commercially and in consumer products. One of the products of interest is an indoor and outdoor caulking compound used for windows, doors, and bathrooms. The primary formulation by the caulking compound manufacturer is enclosed and ventilated, with survey data available. However, no survey data are yet available for use of the product by commercial applicators or consumers. We expect the commercial home repair personnel and consumers to have similar exposures per application, but the commercial applicators’ exposure will be more frequent. Information about consumers’ use of the product suggests that caulking in a bathroom may be the most significant for exposure. Typical consumer use is as follows:
  • exposure duration: 45 minutes
  • frequency: three applications per year (but we will model the exposures for one application)
  • amount of caulking applied: 75 g (most of one hand-squeezable tube)
  • bathroom volume: 10 m3
  • bathroom ventilation rate: a small bathroom ventilation fan of 50 cfm (1.4 m3/min)
Substance X has a molecular weight of 106.17 g/mol and its vapor pressure is 1,280 Pa (9.6 mmHg) at 25 C° (for the pure substance) for a saturated vapor concentration of approximately 13,000 ppm (56,000 mg/m3). The listed maximum fraction of X in the caulking product is 5 percent by weight, which reduces the vapor pressure and the saturated vapor concentration according to the mole fraction and chemical activity, neither of which we currently know. If the mole fraction is similar to the weight fraction and we assume Raoult’s Law behavior, the initial saturated vapor concentration could be 5 percent of 13,000 ppm or 650 ppm (2,800 mg/m3). However, if we model the emission rate from the application rate and weight fraction, we have an emission in mg/min, which is a “standard” form for a generation rate in a model. What form, though, does this emission rate take? Is it continuous or instantaneous? How does our assumption of the emission rate affect the model?
As a first approximation, let’s assume the caulk is applied at a constant rate during the 45-minute exposure durations. The rate of 75 grams applied over 45 minutes is 1.667 grams per minute or 1,667 mg/min, but only 5 percent is X, so the rate of application of X is 83 mg/min. Now let’s use two models familiar to many IHs, the well-mixed room and the two-zone model. Calculations for both of these models are available through IH Mod, a spreadsheet for modeling exposures. (IH Mod can be downloaded as an
Excel document
from AIHA's website.)
In occupational settings, we are often interested in the task duration or full-shift exposure. For this example we’re concerned about exposures to residents and users of the product, so we’ll use IH Mod to run a simulation for a full 24 hours (1,440 minutes).
During application of the caulking product, emissions continue for 45 minutes, then cease, and the space is purged by ventilation. Figure 1 shows the results for the well-mixed room model. Note the rising concentration curve: this is due to the emission rate exceeding the ventilation removal rate during application. The peak concentration at 45 minutes is about 60 mg/m3 with a 45-minute TWA of about 50 mg/m3. The 24-hour TWA is about 1.9 mg/m3.
For the near-field/far-field model, we need to assume a near-field volume and an inter-zonal air exchange rate. Assuming a near-field hemisphere (the bottom of a full sphere blocked by surfaces) of 0.8 m radius (approximately arm’s length) and a relatively low random air velocity of 3 m/min, the inter-zonal air exchange rate is 6 m3/min.
Figure 2 shows the results for the two-zone model. As expected, the near-field results are slightly higher, with a peak concentration at 45 minutes of approximately 73 mg/m3 and a 45-minute TWA of 63 mg/m3. While these results may better represent the applicator’s exposure, the difference from the well-mixed room results is not great due to the significant inter-zonal exchange rate (compared to the room size and general room ventilation rate) and the relatively large fraction of the near field in the small volume of the bathroom. We conclude that the TWA should be based on 45 minutes in the near field and the remainder of the 24 hours in the far field (that is, the other zones of the residence, where the concentration is lower). The result is a slight increase for the 24-hour TWA to about 2 mg/m3.
Below is a list of a few key models. Many others exist. 
The Dutch National Institute for Public Health and the Environment (RIVM):

AIHA: IH Mod (
Excel download

AIHA: IH SkinPerm (
Excel d

Exposure Assessment Tools and Approaches
. •

Consumer Exposure Model


Multi-Chamber Chemical Exposure Model
Stochastic Human Exposure and Dose Simulation

Considerations for Exposure Assessment of Consumer Products