Impact of exposure factor selection on deterministic consumer exposure assessment
Introduction
Exposure to chemicals in consumer products (CPs) is often estimated by indirect assessment using exposure scenarios. Model-based exposure assessments of chemicals in CPs have been conducted using probabilistic and deterministic methods. Probabilistic approaches consider probability distributions in input variables and predict the distribution of exposure in a target population (Cullen and Frey, 1999). Deterministic methods use point estimates of input parameters to provide a single worst-case value (IPCS, 2005). Deterministic methods are often used to screen CPs for hazardous exposures.
Despite the allure of the apparent simplicity of deterministic methods, outcome may be strongly dependent on the selection of the percentile of the distribution of exposure factors. Although exposure factors can be obtained through surveys, behavioral observation, or activity models (Parmar et al., 1997), information on exposure factors is often limited. In order to estimate exposure with exposure factors, knowledge of the type of distribution (ie, normal, log-normal, other) is important, especially when estimating higher ranges of exposures (Hakkinen et al., 1991). However, it is difficult to find the exact patterns of distribution of exposure factors through the small samples. Even though the pattern of distribution on actual population determined by sample, which percentile is chosen as an input value may affect the uncertainty of the results. Deterministic methods can overestimate exposure levels because they use extreme values for the parameters (Fryer et al., 2006). The multiplication of several high percentiles together may result in the unrealistic estimates. In addition, the uncertainty of a result represented by a single value is not quantitatively considered in exposure estimation (Ferrier et al., 2002).
There are some guidelines of selection exposure factors to estimate exposure to CPs by deterministic methods. The Consumer Exposure and Uptake Model (ConsExpo) of the Dutch National Institute for Public Health and The Environment (RIVM) is an example of a CP exposure model (van Veen, 1995). ConsExpo 4.0 recommended that the specific percentile value of each exposure factor's distribution could be used to estimate exposure levels by deterministic approaches (Delmaar and Schuur, 2016, Höglund et al., 2012). To avoid unrealistically high estimates and to maintain conservative exposure estimates, they selected a 75th percentile as a representative value for each parameter (Delmaar and Schuur, 2016). In ConsExpo fact sheets, the 75th percentile of exposure factors (frequency, amount, spray duration, and exposure duration) have been chosen as the default values for CPs such as cosmetics, cleaning products, and disinfectant products (Bremmer et al., 2006, Prud'Homme de Lodder et al., 2006a, Prud'Homme de Lodder et al., 2006b). One study conducted deterministic exposure assessment for CPs using ConsExpo's defaults (Gosens et al., 2014).
The European Center for Ecotoxicology and Toxicology of Chemicals developed the Targeted Risk Assessment (TRA) tool for first tier assessments of consumer exposure (Ecetoc, 2009). Many default exposure factors in TRA were obtained from the RIVM fact sheets. When specific information was not available, values were derived using expert judgment. The United States Environmental Protection Agency (US EPA) developed the Consumer Exposure Model (CEM) for cleaning products (EPA, 2017). CEM provided the exposure factors by three classes of high, medium and low.
Although a certain percentile was proposed in the selection the exposure factor in deterministic assessment the precise degree of uncertainty remained unknown. Errors associated with the selection of a specific percentile for an exposure factor need to be determined. The purpose of this study was to compare inhalation exposures estimated by a specific percentile of each of the three exposure factors in deterministic assessment with population exposure. Population exposure was calculated by exposure factors of 3333 participants representing the national population.
Section snippets
Methods
This study utilized use patterns of 9 CPs collected from 3333 participants. In a previous study, exposure factor data for 9 CPs such as a deodorizer (fabric deodorizer), cleaning products (dishwashing detergent, bathroom cleaner (bottle and trigger type), toilet rim cleaner, glass cleaner, and floor cleaner), and disinfectants (household bleach and mold stain remover) were collected (KNIER, 2012). Detailed information pertaining to data collection has been reported elsewhere (Park et al., 2015
Exposure factors of CPs in the parent population
The exposure factors of the 9 CPs were obtained from 3333 participants. The number of users, medians, and ranges of the exposure factors in the parent population are shown in Table 3. The most frequently used product was dishwashing detergent (median = 2/day). The median frequency use of glass cleaner was the lowest (approximately 2/month). The median amount used per application was highest for bottled bathroom cleaner (70.25 g/event) followed by toilet rim cleaner and floor cleaner (both
Conclusions
Inhalation exposure by deterministic methods could have errors due to selection of percentile of exposure factors. Inhalation exposures by the 75th percentiles of each of the three exposure factors were much lower than the 95th percentiles of the parent population exposures. For 9 CPs, exposures using a range between the 85th to 99th percentiles of each of the three exposure factors were closer to the 95th percentiles of the exposures of the parent population. Based on our findings, the
Acknowledgements
This study is supported by the Korea Ministry of Environment as “The Environmental Health Action Program” (#2015001940002).
References (19)
- et al.
Human exposure modelling for chemical risk assessment: a review of current approaches and research and policy implications
Environ. Sci. Pol.
(2006) - et al.
Determining the exposure factors of personal and home care products for exposure assessment
Food Chem. Toxicol.
(2015) - et al.
Stepwise approaches for estimating the intakes of chemicals in food
Regul. Toxicol. Pharmacol.
(1997) - et al.
Cosmetics Fact Sheet. To Assess the Risks for the Consumer
(2006) - et al.
Probabilistic Techniques in Exposure Assessment: a Handbook for Dealing with Variability and Uncertainty in Models and Inputs
(1999) - et al.
ConsExpo Web: Consumer Exposure Models-model Documentation
(2016) TR 107 : Addendum to ECETOC Targeted Risk Assessment Technical Report No. 93
(2009)Consumer Exposure Model (CEM) Version 2.0 User's Guide
(2017)- et al.
Current knowledge and recent developments in consumer exposure assessment of pesticides: a UK perspective
Food Addit. Contam.
(2002)