Conclusions

While being a well-established component of risk assessment, exposure assessment is still an often-overlooked step in establishing the safety of a chemical. Greater emphasis needs to be placed upon exposure assessment if purely hazard-driven approaches are to be avoided in the safety assessment of chemicals. This is particularly the case in light of the current trend that considers alternative techniques like in vitro assays and in silico models for establishing a safe dose of a chemical for humans, which are becoming more and more common place as we move towards alternatives to animal testing. Such methods are used as both screening techniques for the prioritisation of chemicals for risk assessment and as a complete alternative to animal testing to establish safe doses protective of human health. Also, it should always be borne in mind that once a safe level or health based guidance value is established for a chemical, the exposure aspect is often the only avenue available to the risk assessor and risk manager in terms of analysing and affecting human exposure and therefore safety.

Despite the importance of exposure assessment, challenges still remain and have been highlighted in this report. The first is access to the appropriate tools and data for conducting consumer exposure assessment. While many sources of information are available, no centralised repository exists where all sources are catalogued for use, detailing domain of applicability and the nature of the data. Additionally, it is not always clear what level of detail or what tier of exposure assessment various data sources and tools are appropriate for. Some efforts have been made to create a resource providing an overview of available tools and data sources for consumer exposure assessment, as well as detailing the nature of the data and tools. While this represents an important first step, it should be recognised that the resource is not exhaustive. A resource of available tools and data sources for consumer exposure assessment should always be a dynamic resource that is continually updated, ideally housed in a web-based platform that can be interactively accessed and refreshed as new tools and data sources become available.

In terms of gaps identified for data sources and tools, key amongst these were a lack of a database on the chemical composition of consumer products based on actual use levels. Such a resource is important for developing realistic exposure estimates, as it is often the first port of call when refining an exposure assessment which is often based on maximum authorised use levels. Such a database often raises the issue of access to proprietary data; however specific use levels are not always required as ranges or estimates of statistical distributions of concentration are often all that is required. Similarly, resources that determine the frequency of occurrence of a chemical in consumer products or the chemical occurrence (often derived from market survey databases) are a key resource that should be greater developed owing to the refinement they offer in an exposure assessment.

Knowing then what data and tools are available, an immediate question posed is what is the appropriate use of different data for different types of exposure assessments? This question was addressed by examining the nature of different types of data and tools and via two case studies in aggregate exposure. It was established that low tier tools are not appropriate for determining aggregate exposure, due to their inherent conservatism. This is particularly true for tools and data sources for industrial chemicals in consumer products outside of the cosmetics and personal care products domain. Should aggregate exposure be required for that area, greater work and effort is required to gather the appropriate data on product co-use. What may be of greater benefit is to place resources into techniques that establish whether aggregate exposure is required at all. For example, the case studies indicated that the additional resources dedicated to aggregate exposure predictions could result in exposures similar to maximum screening level predictions for single products. For domains where higher-tier tools are available (namely cosmetics), the considerable refinements that can be introduced by data on consumer habits and practices, product co-use, refined concentration values and presence probabilities was demonstrated using the examples of triclosan and phenoxyethanol. These case studies re-establish an important point; if the margin of safety initially established in a screening level exposure assessment is unacceptable, the next steps could be either a more refined exposure assessment or a risk management measure. Factors such as uncertainty, variability, and initial skew or bias in the original estimation should be considered to assist in decision making as to which of these steps is appropriate.