Risk Assessment
John Dege, DuPont Representing the chemical industry’s Acrylonitrile Group
Controlling Air Toxics in a Post-MACT Environment
• We need to recognize how far we have come, with:
– 20 years of state air toxics programs focused primarily on major stationary sources. – 10+ years of MACT promulgations for major stationary sources.
• Post-MACT stationary source emissions are now relatively small and potentially very costly to control further.
1999 NATA Cancer Risk
12%
42%
17%
Major Area/Other On-road Non-road Background
Source: Ted Palma, OAQPS
21%
April 15, 2004
8%
DRAFT results
In this context-• We need to determine which types of sources pose real risk problems and change our programs accordingly. • Any new controls need to be carefully targeted to individual facilities/units that still pose significant risks. • Any new regulations need to provide these sources flexibility in how they comply.
The Key Role of Risk Assessments
• Identification of the real risk problems and tailoring our controls to those sources will require potentially thousands of carefully prepared RISK ASSESSMENTS. • Unfortunately, our health value database and exposure inventories fall far short of what we really need for these risk assessments.
Health Values
• Our Goal should be: Use the best available health effect information for regulation. • Challenge: We have a less than up-to-date set of health values for hazardous air pollutants.
The case of acrylonitrile
• Background:
– Current IRIS cancer value based on an outdated (1980) epidemiological study. – EPA identified what further research was needed. – Industry undertook to do those studies. – The resulting database is one of the best human data sets available for any industrial chemical. – Recently, a hazard assessment based on the complete database was independently peer reviewed by a TERA panel.
Acrylonitrile health values
• This new risk assessment results in numbers significantly different from the old IRIS values. • Why?
– IRIS scientific process more sophisticated today. – Epidemiology cohorts were expanded and updated, and original effects of concern were shown to be an artifact.
Acrylonitrile health values—cont.
• What do we mean by ―significantly different numbers?‖
– AN would no longer qualify as a risk driver in EPA screening assessments for residual risk (10-6 and H.I. of 1.0). – There is little or no threat to public health from environmental exposures to AN. THIS IS REALLY GOOD PUBLIC HEALTH NEWS!
What are the lessons learned?
• As air pollution professionals, when we hear credible evidence that an IRIS value does not reflect current science, we need to respond accordingly.
– We need to keep in mind that the average IRIS values are over 10 years old.
• Use of an outdated value could mean that the chemical may be under- or over-regulated. • A case-by-case approach may have resource implications, but the newer science should not be ignored.
The IRIS database gives us clues of when we NEED to look further.
A common IRIS entry:
―Screening-Level Literature Review Findings -- A screening-level review conducted by an EPA contractor of the more recent toxicology literature pertinent to the cancer assessment for [chemical] conducted in September 2002 identified one or more significant new studies. IRIS users may request the references for those studies from the IRIS Hotline at hotline.iris@epa.gov or (202)566-1676.‖
OAQPS’s Health Value Look-up Tables Need to Reflect the Latest Science
• http://www.epa.gov/ttn/atw/toxsource/summary.html
Up-to-date health values should be used in:
• National and regional assessments (e.g., NATA) • Risk assessments that support rules (e.g., Residual Risk) • Site-specific risk assessments to demonstrate compliance • Comparable state program assessments
Exposure
• Hazard assessment often gets most of the resources and attention. • Our Exposure assessment tools (e.g., emission inventories) have suffered as a result.
Exposure
Hazard Assessment
Assessment
National Emission Inventory
• We have a largely ―hit or miss‖ national emissions inventory for hazardous air pollutants, despite considerable effort expended in the past couple of years. • We found that even a little double-counting of facilities in a single county can have a dramatic effect on the perceived national risk from a chemical such as acrylonitrile. • The challenge: What should we as air pollution professionals do when we know or suspect that our exposure data are wrong?
Example: Acrylonitrile Emissions from Landfills
• 1996 NATA: 39% of AN emissions come from municipal landfills. • Calculated by: emission factor X every municipal landfill in the country. • Emission factor is based on 4 old measurements of landfills that likely contained industrial waste. • New emission data from landfills now being collected (slowly). We expect results to show AN emissions from landfills are zero or close to it.
What is everyone’s responsibility in the meantime?
• When we know or suspect emission factors or other exposure data are wrong, what should everyone do?
– Option #1: Continue to use the old data and hope no one notices. – Option #2: Use the old data, but caveat it prominently. – Option #3: Seek out new estimates based on best available science.
Our recommendation
• Taking the easy course—option 1—of ignoring the problem seems the most popular choice today. It’s cheaper. • However, options 2 and 3 are more professionally responsible, more protective of public health, and more respectful of not wasting resources on inappropriate regulations.
Summary
• Collectively we need to:
– Work to upgrade the health and exposure data systems that support risk assessment, and – Using the best available science in our risk assessments, identify the sources of the real remaining risks, and tailor our programs and controls accordingly.
• AND…if we have to use shortcuts in these assessments, we should include prominent caveats for our conclusions, and invite those who might have better data to come forward with it for us to consider.