The health risk assessment methods used by Scorecard closely correspond to the risk assessment methods used by federal and state agencies in carrying out regulatory duties. Such methods involve a number of important uncertainties. These risk estimates assume that a large population with a standard distribution of age and other demographic factors is exposed for a lifetime. Because the actual demographic features and exposure duration for a particular population may vary significantly from these assumptions, it is not appropriate to use these figures to make projections of the numbers of additional cancer cases that a particular local population will experience. Scorecard's risk estimates provide a screening-level perspective on the potential magnitude of public health impacts from a facility, which are useful for comparative purposes. They are not necessarily predictive of any one actual person's risk of getting cancer.

LIMITATIONS OF EPA'S EXPOSURE ESTIMATES
EPA's exposure estimates have important limitations:
1) They are generated using a computer model, rather than using actual measurements taken in the area of the facility.
2) Because of modeling limitations, some chemicals that are released cannot be included in risk estimates.
3) EPA assumes that exposures to its estimated concentrations of chemicals are continuous, and are at the same level indoors and outdoors.
4) EPA's estimates are based on direct inhalation only and do not include any exposures from water or food that has been contaminated from chemicals in the air.

EPA'S RISK-SCREENING ENVIRONMENTAL INDICATORS PROJECT
EPA's Environmental Indicators project is designed to conduct screening-level analysis of risk-related impacts of toxic chemical releases in the U.S. To estimate exposures to chemicals released into the air by TRI facilities, a sophisticated software package combines a facility's reported TRI data with information needed to model resulting ambient air exposures to the surrounding population. (The software has recently been extended to model exposures resulting from water releases.) After the Environmental Indicators model estimates the chemical doses that individuals in an area are predicted to experience, EPA combines this exposure information with population data and toxicity weights to produce a risk-screening score. EPA's score for a facility does not indicate a cancer risk estimate. See EPA's discussion of the strengths and weaknesses of its scoring system.

NOTE: Scorecard uses only one part of the information generated by the EPA Risk-Screening Environmental Indicators project: EPA estimates of exposures to reported chemical releases from TRI facilities. Scorecard then calculates its own potential cancer risk estimates as described below. Scorecard does not report EPA's risk-screening scores.

EPA's Environmental Indicators project uses the same approach as a standard air dispersion model (the Industrial Source Complex Long Term (ISCLT3) model) to estimate long-term pollutant concentrations surrounding a stack or area source. Predicted ambient concentrations are a function of facility-specific parameters such as stack height and exit gas velocity, local meteorology, and chemical-specific first-order air decay rates. Concentrations are predicted by ISCLT3 at the midpoint of each 1 km by 1 km grid cell in a 100 km by 100 km area surrounding the facility. EPA has conducted analyses that show that concentrations predicted by the Indicators model using a combination of generic and site-specific data closely match concentrations estimated by using more complete site-specific data.

For each grid cell in the modeled area for a facility, the predicted concentration of a chemical is combined with standard human exposure assumptions to estimate a "surrogate dose." EPA then multiplies this dose by the population in the cell and by the inhalation toxicity weight assigned to a chemical to derive a risk-related score. EPA's overall Environmental Indicator score for onsite air releases from a facility is calculated by summing chemical-specific scores over all cells affected by a facility's release.

The Risk-Screening Environmental Indicators software system is distributed by EPA for free and allows users to conduct their own risk-screening analyses of TRI facilities. Scorecard uses Version 2.0 of the RSEI model.

HOW SCORECARD OBTAINED EXPOSURE INFORMATION FROM EPA'S ENVIRONMENTAL INDICATORS PROJECT
Scorecard derived chemical-specific "surrogate dose" data for each facility releasing recognized or suspected carcinogens by dividing EPA's chemical-specific full model score for a facility (obtained from the from the Risk-Screening Indicators software) by EPA's inhalation toxicity weight for that chemical and by the total population in the modeled area for a facility. Unfortunately, the design of EPA's software does not allow for a corresponding derivation of the dose data required for noncancer health risk assessment.

HOW SCORECARD PRODUCED CANCER RISK ESTIMATES USING EPA'S EXPOSURE INFORMATION
The chemical-specific "surrogate dose" derived from EPA's Risk-Screening Indicators project was multiplied by an inhalation cancer potency to yield an upper bound estimate of an average individual's added cancer risk from exposure to that chemical. Cancer risks for each chemical were assumed to be additive and were summed together to derive a total estimate of added cancer risk from the air emissions of a facility. Only chemicals that Scorecard identifies as recognized or suspected carcinogens were included.

Because Scorecard divides EPA's full model score by the total population in all cells of the modeled area for a facility and not by the exposed population (smaller than or equal to total population), Scorecard may underestimate the average individual's added cancer risk for some facilities.

ESTIMATED ANNUAL AVERAGE CONCENTRATION
Scorecard also estimates the annual average ambient air concentration of each chemical. It does so by converting EPA's estimate of the average surrogate dose of the chemical received by the population around a facility into units of concentration. Standard exposure assessment assumptions were used in the conversion.

Because Scorecard divided EPA's full model score by total population in the modeled area and not by the exposed population (smaller than or equal to total population), the annual average concentration may be underestimated for some facilities.


REFERENCES

EPA. Risk-Screening Environmental Indicators Model: Version 1.0 [1988-1997 TRI Data, Air-Only Model]. July 6, 1999.

EPA. User's Manual for OPPT's Risk-Screening Environmental Indicators Model: Version 1.0 [1988-1997 TRI Data, Air-Only Model]. July 6, 1999.

EPA. Risk-Screening Environmental Indicators Methodology.