skip to main content

→ Top Stories:
Clean Power plan
Safe Chemicals

Jennifer Sass’s Blog

National Academies heaps praise on the EPA IRIS chemical assessment program

Jennifer Sass

Posted May 6, 2014

, , , , ,
Share | | |

The EPA's program to assess the hazards of chemicals - called the Integrated Risk Information System (IRIS) - earned itself a lot of praise from the prestigous National Academies (NAS) for its adherence to scientific excellence, and to public transparency and accountability. 

Today's National Academies report (here) comes after a robust, two-year long review of the IRIS program by fifteen scientific and technical experts and numerous peer reviewers. It was initiated in response to a 2011 National Academies report that, although focused on an IRIS formaldehyde assessment, provided many general recommendations for improving the IRIS approach, some specific methods, and its process. That formaldehyde report also provided a roadmap for the IRIS program to move forward. 

While I share the National Academies' overall praise and enthusiasm for the improved IRIS process, there are several specific comments from the report that are particularly pleasing:

  • The report supports "firm stopping rules for key points throughout the process to guard against delay..." (page 9) and "enhance productivity" (page 13, 24). These stopping rules should help prevent industry tactics that delay chemical assessments (See our Delay Game report for documented examples).
  • Although industry routinely touts Good Laboratory Practices (GLP) - which are required for many industry studies - as the most important measure of a good study, the report noted that they fail to address study bias (page 63). In fact, GLP standards were imposed on industry labs during the 1970’s after flagrant violations and fraud were identified.
  • The report notes that bias is a systematic, not random, flaw in the design and conduct of a study that reduces the validity and reliability of the study results  (page 62, 63). Industry-sponsored studies have frequently been shown to be "biased by design" to underestimate risks (see Doubt Is Their Product for documented examples of industry biased studies).
  • The report recommends following rules such as those used by the International Agency for Research on Cancer (IARC) that prevent experts with conflicts of interest from participating on scientific review panels (page 92). I hope the IRIS program takes that recommendation seriously! 

The report specifically praises the leadership - Drs. Kenneth Olden and Vincent Cogliano - for the improvements, and for their successful efforts to gather input from stakeholders and the general public (page 129), despite overwhelmingly uneven participation by industry and its representatives (page 22).

NRDC also commends Drs. Olden and Cogliano, and their top-notch staff, for continuing to provide high quality scientific assessments - despite unrelenting industry opposition - of the toxic chemicals that are in our food, water, air, and consumer products, so that health-protective steps can be taken to remove, reduce, and replace them.


Share | | |


Tony TweedaleMay 6 2014 05:50 PM

Good on NAS/NRC! Which bias does NRC say GLP studies have? I would guess a bias away from sensitivity (the ability to detect toxicities).

GLP is not the problem (as you say it exists due to the IBT industry lab fraud). The problem with these toxicity tests are the other mandate that laws globally require pre0-market risk assessment to adhere to: the OECD Test Guidelines (US gives them its own names, but they are identical). These test methods originated in industry labs and are grossly insensitive to find toxicity. What does the NAS say about TGs?

Patrick CoffeyMay 7 2014 12:01 AM

You mention Doubt Is Their Product, a great book, but I thought the author, David Michaels, was skeptical of IRIS, because it is a computer simulation. Am I right, or am I thinking of something else?

Jen SassMay 7 2014 11:05 AM

Dear Patrick and Tony - thanks for your comments. Tony, i'm not sure what you are thinking of, but the IRIS program reviews laboratory toxicity studies, epidemiologic studies, molecular/mechanistic information, and any other information that is available and relevant.

Tony TweedaleMay 8 2014 08:04 AM

Thanks Jen, and I'll clarify(!). I was seeing if you would expand on what the NRC cmttee. had to say about bias being systematic (controllable, not random).

Also I said that the biases they note about industry GLP tests come from the OECD TG test methods, not from GLP. Further, that the main bias (away from sensitivity) is due to testing chronic doses that are so high they really are the 'end of poisoning' and so deliver L/NOAELs needed to do a RA...but do not test at all what effects realistic doses might be having (as thousands of published findings of such accumulate, but every one is barred from pre-market RA).

A related big insensitivity is they "incinerate the evidence"--sacrifice animal as soon as dosing ends, not letting any chronic disease induced develop (over half of cancers develop in humans after the equiv. age the dosing stops in a rodent 2 yr test!). Diddling control animals is another bias.

Comments are closed for this post.


Switchboard is the staff blog of the Natural Resources Defense Council, the nation’s most effective environmental group. For more about our work, including in-depth policy documents, action alerts and ways you can contribute, visit

Feeds: Jennifer Sass’s blog

Feeds: Stay Plugged In