How Are Chemical Harm Risks Established After an Industrial Spill?

I understand that this is the first time the EPA has formally put together a document like this?

Correct. This started back in 2014. The EPA has been strengthening their processes over the years to develop this first handbook to describe what their systematic process is for reviewing chemicals and determining if they are a hazard or not. 

How is this new IRIS approach different?

It’s a really different approach by the EPA. They used to always do a very broad search to look for all the animal and human evidence, but it wasn't always a systematic search. They didn't assess the quality or relevancy of the evidence and they didn't try to synthesize it. With this new approach there will be a reduction in bias in the assessments of chemicals. 

What were some of the committee’s recommendations to improve the EPA’s handbook?

Our bottom-line assessment of the handbook was that it was a great first effort and that it actually should serve as a systematic model for other areas of EPA as well as other agencies that do risk assessment, such as under the Toxic Substances Control Act. 

One major area that we commented on was what's called “risk of bias” or “quality assessment.” EPA had suggested methods for doing quality assessment of both animal and human studies that were a bit inconsistent. For example, one area was around who is funding the studies the EPA is considering. In our comments, we demonstrated that the results of studies are associated with who funds them regardless of all the other quality criteria. It doesn't mean they should exclude industry funded studies, but the EPA should  see how the results differ and weigh them less heavily in the final analysis. 

The other big area was how the EPA establishes the level at which something creates a hazard. There's a disconnect in the handbook. It wasn't clear how the studies from the systematic review flowed into the selection of the studies for the toxicity level. And this is something EPA is very sensitive to because in their older methods, they used to basically pick one study to establish the toxicity level. So they may have reviewed like 80 studies and then they'd say, ‘Oh, this one looks good, and we'll use that for the toxicity level.’ And there wasn't any really clear rationale for why they picked that study. And so it was better in the handbook, but still we think that the systematic review should flow directly into the selection of the studies for the toxicity level.

When you talk about establishing risk for chemicals, is that risk level set for healthy, full-grown adults?

That is a great question because one of the comments we had in the handbook for an area that hasn't really been adequately addressed is how the EPA is going to handle analysis of what's commonly called “sensitive populations” and “developmental stages” – prenatal, elderly, everything in between.

Right now, the EPA, when they do an assessment of a chemical, they look very broadly at all the literature and studies done on a particular chemical. But if the majority of the studies have been done in healthy adults, that could be their main pool of evidence. And the EPA may actually just park the other studies of sensitive populations and subgroups for later analysis. For example, if we want to know how a chemical might affect pregnant women who might be working in a factory where they can be exposed, then that would include that sensitive population. 

It depends on the question that's put to the EPA for each chemical. But in our report, we noted that the handbook didn't have a good or a clear method for saying how that sensitive population should always be addressed and how they should be addressed.

How does the IRIS handbook get deployed in the field after an industrial spill?

It'll depend on what the chemical is and if there already has been an IRIS assessment conducted. 

One thing worth noting: IRIS assessments take some time to conduct – considering peer review and through revisions in the methodology. 

In the case of a chemical spill – say it's chemical arsenic – then the IRIS assessment would be referred to determine: Is this chemical a hazard? And if so, at what level is it a hazard? 

In determining levels, the EPA also takes into account the route of exposure – is it in the air or water? Is the arsenic impregnated in wood that you can get chemical dermal touch exposure?  

What the IRIS assessment determines is basically whether the chemical is hazardous. From there, other parts of EPA that would say, "OK, arsenic in the water, we know what the level has to be, but how do we keep it at that level? What needs to be done? Can we get it out of the water? If it's in the water already, how do we control that level?" 

So what's the next step? Is the EPA going to go through another subsequent review of everything and then publish the final IRIS after that?

Yes. They've published their next draft already. The EPA responded really well to feedback implementing the advice the committee offered into the next draft of the handbook, but there’s still improvements to be made. 

I don't think they'll ever call it final, because part of the recommendation was that this is an iterative  process. It's going to improve. The methods are evolving.