In the early 1980’s
EPA contracted with the Research Triangle
Institute (RTI) to develop a “Tolerance
Assessment System” for use in computing
human exposure to pesticides in foods consumed
in the United States. (The Tolerance Assessment
System was later named the “Dietary
Risk Evaluation System” (DRES), and
this latter name will be used in the remainder
of this report.) The food intake data used
in the original DRES was based on 3-day
diaries of food intake reported by 30,770
individuals in the 1977-78 Continuing Survey
of Food Intakes by Individuals (CSFII) conducted
by the USDA.
Since the foods reported as
consumed in the CSFII were “as eaten”
they had to be broken down first to their
“raw agricultural commodity”
(RAC) ingredients and then those ingredients
individually broken down further to specific
“food forms” (e.g., raw, frozen,
cooked, baked, boiled, etc.). RACs and foodforms
(FF) were assigned numerical codes and “recipes”
were developed which converted foods-as-eaten
to their RAC/FF constituents on a percentage
basis. Daily intake amounts by RAC and FF
were then computed for each individual on
each of the 3 days of the CSFII survey and
stored on tape. Mainframe computer programs
were written to create this database of
individual intake amounts and to process
this database with real or hypothetical
chemical residue amounts (“tolerances”)
in order to estimate the distribution of
residue intakes in the U.S. population and
specific subpopulations (e.g,. infants,
children 1-6, etc.). “Chronic”
exposure computations (mg/kg-body weight/day)
for the U.S. population and specific subpopulations
were computed by multiplying mean daily
RAC/FF intake amounts (adjusted for individual
body weights) for the entire population
or subpopulation by RAC/FF-specific residue
amounts and summing the exposure amounts.
Since the chronic estimates
were based on mean intake amounts, no distribution
of intakes or exposure amounts could be
determined. “Acute” exposure
computations (mg/kg-body weight/day) were
based on individual daily RAC/FF intake
amounts (adjusted for body weight) multiplied
by RAC/FF-specific residue amounts, with
daily exposure amounts computed for each
day separately for each individual in the
CSFII subpopulation, and then used to build
a distribution of daily exposure amounts.
Statistical weighting factors provided by
USDA for each participant in the 1977 CSFII
were used to weight the exposure amount
for each individual in computing both the
chronic and acute estimates of exposure.
Margins of safety based on chronic and acute
reference doses, or NOELs (no-effect levels),
could be computed, and in the case of acute
exposure computations, the percent of the
subpopulation that might be exposed to unacceptable
intake amounts could be estimated.
DRES dramatically improved
the scientific basis for setting crop-specific
tolerances for pesticides. Now actual combinations
of foods consumed by individuals in the
U.S. could be used for estimating total
exposure, as opposed to the “worst-case”
assumption that all foods with pesticide
residues of interest were consumed on the
same day at some arbitrarily high level.
DRES’s chief limitation was that it
required a main-frame computer and computer
programmers to perform even simple calculations,
with slow turn around times and little flexibility
for making computational or reporting changes.
Dr.
Barbara Petersen played an important
role in the development of DRES as an independent
contractor hired by RTI, especially in translating
the foods-as-eaten reported in the CSFII
to RACs (raw agricultural commodities) used
in pesticide residue reporting. After completing
the RTI contract, Dr. Christine Chaisson
joined her consulting practice and it was
renamed “Technical Assessment Systems,”
with a mission to provide professional expertise
related to food safety issues, especially
issues related to pesticide residues in
foods. Dr. Petersen told Steve Petersen,
at the time an economist at the National
Bureau of Standards, of her need for software
that could run on a microcomputer (still
in its infancy at that time) and reproduce
the chronic exposure calculations of DRES
to better understand the sensitivity of
these calculations to a variety of “what-if”
scenarios. (At that time the idea of developing
an acute exposure analysis program that
could run on a microcomputer was not even
considered because the database of individual
RAC/FF intake amounts by day based on the
entire 1977 CSFII was too large for a desktop
PC in the days before a hard disk with adequate
data storage capacity was available and
affordable.)
