Confused by All the Tools to Evaluate Chemicals?

While listening to the radio in New York City a while back I heard an ad for a ‘chemical free’ mattress.What is a ‘chemical free’ mattress? There is no such thing. In fact, like all of our consumer products, all humans, animals, and minerals are made up of chemical molecules. Some molecules are made by nature, think wood or cotton or biomaterials (grass, bamboo), and others are made synthetically (or by man). By taking what nature has given us, companies turn biomaterials, oil, natural gas, salt, minerals, etc. into the beneficial and innovative products we use today. There is no way to avoid chemicals.There are no ‘chemical free’ products, period. Since chemical molecules are part of everything, we need to understand them.  So how do we go about this process? 

One popular way is through the use of chemical evaluation tools. Many of these tools are online resources or automated tools provided by companies. There are literally hundreds of tools and databases available in the marketplace, so how does one sort through these tools to determine which are the most thorough?; the easiest to use?; the most cost effective?; look at the most toxicological endpoints?

Last week in the journal Integrated Assessment and Environmental Management (Gauthier, A.M. et al., 2014… I am one of the et al’s) we published a paper called “Chemical Assessment State of the Science: Evaluation of 32 decision-support tools used to screen and prioritize chemicals.” Knowing that most of you are not into reading obscure toxicology peer reviewed journal articles, this article summarizes the key top-line findings from this important paper.

What did we do? In this paper we evaluated common tools, each of which varied with respect to their intended use and purpose, using a series of independent evaluative criteria grouped into five categories:

1) Screening & Prioritization (S)

2) Database (D)

3) Hazard Assessment (H)

4) Exposure & Risk Assessment (E)

5) Certification & Labeling (C)

Each tool was evaluated against a number of criteria to determine its robustness and usability.  100 points were assigned based on how the tool performed on the following criteria:

• Hazard or Hazard & Exposure – A greater number of points were given to tools that evaluated exposure in addition to hazard, since this gets to the real risk of a product throughout its life cycle.
• Hazard Endpoints Considered – More endpoints, equals a more thorough evaluation and thus achieved more points.
• List-based verses Criteria-based – Setting a criteria for what constitutes a chemicals that poses a risk gets more points than simply relying on a list. Lists tend to be backward looking and thus no new chemicals will be on a list, but it might meet the criteria for further evaluation.
• Data Gap Provisions – How does the tool deal with less than perfect data, if at all. 
• Chemicals Available for Assessment – Points were given to tools that have a large list of chemistries available in their database or that can be evaluated without generating primary data.
• Review Process – How are decision made to change or update the tool?  Is it two guys in a room or is it an open, multi-stakeholder process?
• Transparency – Is the tool a black box or is it an open box in terms of how results are achieved?
• Ease of Use – Do you need a PhD in toxicology to work with the tool or can a person with basic scientific skills achieve acceptable results?
• Accessibility / Cost – Is the tool free and easily accessible or does it take thousands of dollars and months of labor to achieve results?
• Exposure Parameters included – Are exposure scenarios captured in the output and if so, how robust are they?

Discussion

First, it is important to understand there was no ‘perfect’ tool based on our analysis.  Some tools scored far better than others, but even the best tools only scored in the 70’s out of 100 possible points.  Our analysis identified that only four of the 30 tools received the maximum score of 10 for assessing both hazard and exposure.  Also, 21 of the tools used a criteria or data-based assessment, but nine relied solely on a list of chemicals.

Most of the tools did not use exposure or the function of the chemical molecule; instead they focused only on the hazardous nature of the chemical molecule in the material.  We know from experience this can lead to poor product choices and regrettable substitution.  For example, preservatives and biocides are added to many products, especially moist or liquid products like those used in personal care products, to play an important, even critical role, in minimizing biological growth.  Since bacteria and germs are part of our natural world and cause adverse, even serious impacts on humans, it is essential that chemical molecules like preservatives are there to inhibit and kill these natural but potentially harmful organisms.  To do their job preservatives and biocides must be biologically active (code for must be able to kill or inhibit the natural growth of microorganisms in a product).  This means they are likely to score poorly on a hazard screen that does not evaluate the functional reason for chemical molecules in products.  Getting rid of preservatives or reducing them to too small quantities does have unintended consequences.  As an unfortunate example, there is a documented case where infants were seriously sickened and one died when the shampoo used by a hospital had its preservative minimized to reduce the hazardous components.  This one example highlights the need for the move away from using just a hazard-only approach. 

Many of the tools scored poorly on how they dealt with the lack of data.  Some tools automatically gave the lowest score; some evaluated the nature of the missing data, e.g. measurement of vapor pressure would not apply for a mineral or solid, because minerals and solids usually do not have a vapor pressure; some simply ignored the lack of data. 

In a sense this is about transparency.  How transparent is the tool provider to disclose the assumptions being made when faced with limited or poor data.  Also, transparency is a factor in the review process for many tools.  Again this was an area where many tools scored poorly.

The good news is that there is lots of competition in this space.  We are already seeing more tool developers looking for ways to include exposure attributes into their systems.  We are also seeing the purchasers of the tools or services of tool providers beginning to ask tougher questions as they seek to make better decisions.  This all leads to better products in the long run.

Next

Next up, we plan to take some of the leading tools and run specific chemicals through them to see if they reach similar outcomes.  Based on the initial results, you’ll be surprised how different the answers are.  Stay tuned.

My thanks to my co-authors who helped in the evaluation of all these tools, Gauthier, A., Fung, M., Panko, J., Perez, A., Hitchcock, K., Ferracini, T., Sahmel, J., Banducci, A., Jacobsen, M., Abelmann, A., Shay, E.