This month, the International Agency for Research on Cancer will issue another paper suggesting a chemical causes cancer - probably one of the compounds in coffee - and journalists will read what IARC actually claims about calculating risk and assume IARC calculates risk. Then, after blowback from scientists who do not consult for Environmental Defense Fund or were not hand-picked to be on the IARC committee, IARC will state they don't talk about risk.

There is a difference between hazard and risk, of course. Caffeine is far more hazardous than BPA, glyphosate or aspartame but, like with those three compounds, you'd have to drink 7,000 cups of something containing them per day to get a toxic effect.

Yet IARC's findings are often strange, and that may be because even tools for finding hazard are not equal. A comparative evaluation of five hazard screening tools published in Environmental Assessment and Management journal found that they gave different results. How seriously can stakeholders, including consumers, retailers and product manufacturers, take results when they vary based on the tool? 

How can you know that chemical is really "greener" if the green company marketing it just used results from the tool that gave the best result? The various approaches analyzed in the study focus on a chemical’s innate “hazard,” meaning they consider whether the inherent properties of a chemical substance could cause harm to humans or the environment under any circumstance. Hazard-based screening does not consider how a chemical is actually used in a product, how much of the chemical substance exists in those uses, and whether and to what degree there is human or environmental exposure to the chemical substance through such uses. These properties of use, dose and exposure are a key component of regulatory chemical safety assessments conducted by government agencies around the world.

The study reviewed select chemicals using the following hazard-based screening tools, to evaluate each one’s process and criteria for assessing a chemical’s hazard profile and determine if consistent outputs resulted from all tools:

GreenScreen Full Assessment
U.S. EPA Design for the Environment/Safer Choice
GreenSuite
SciVera Lens
GreenWERCs (run in four modes: GreenWERCS GreenScreen List Translator, GreenWERCs Green Screen Scoring Model, GreenWERCs Walmart Scoring Model, and a user-defined GreenWERCS ChemRisk Model)

The seven chemicals evaluated included two natural compounds (caffeine and citric acid), a degradation metabolite (glycolic acid), and four synthetic chemicals: ethylene glycol, dibutyl phthalate, benziothiazolinone (BIT), which is also an antimicrobial, and 1,2,4,6,9,10-hexabromocyclododecane (HBDC), which is also a chemical designated as persistent and bioaccumulative by the European Union.

In short, to judge a product’s safety, a chemical hazard analysis needs to be put into the context of how the product is used, and by whom.

The results show a wide variety of hazard categorizations for the same chemical—ranging from little or no hazard/toxicity to very high hazard/toxicity. For example, caffeine was ranked by two of the tools as a “low” hazard, as a “moderate hazard by one tool, as a “high” hazard by three tools, and as a “very high” hazard by two of the tools.

The authors indicated that the different results for the same chemicals were due to variations in: 1) the endpoints each tool considered for evaluation; 2) how each tool weighed the relevance of specific endpoints; 3) the sources of information the tools developer used to gather information; and 4) how each tool treated gaps in available data. Thus, the study found that the outcome of a hazard tool assessment is highly dependent on the tool selected for the screening.