The First Microcomputer-based
Dietary Exposure Programs
Stephen Petersen developed
a prototype system, written in QuickBASIC
and running on a KayPro microcomputer (with
the “CPM” operating system)
for calculating chronic exposures to pesticides
in user-specified foods that would run quickly
and could be used fairly easily by a non-programmer.
This chronic analysis program was based
on pre-calculated mean per-capita intake
amounts for each RAC and FF in DRES for
the U.S. population and pre-selected subpopulations
of interest. A dietary residue file format
was developed for use with the chronic analysis
program, and a residue file editor was developed
that allowed a non-programmer to set up
a list of RACs and foodforms with deterministic
residue amounts and adjustment factors (to
adjust exposure calculations for percent
of crop treated, etc.) for the exposure
analysis. This program was soon converted
to a format for use with the MS-DOS operating
system, named “Exposure 1TM”,
copyrighted in 1985 by the author and licensed
exclusively to Technical Assessment Systems.
Technical Assessment Systems demonstrated
the program to EPA, performed extensive
QA (Quality Assurance), showed that its
results were identical to the chronic calculations
from the EPA DRES, and began using it as
a business tool, licensing it to commercial
end users.
In 1988, when hard disks that
could store more than 50MB of data became
available for desktop computers, Stephen
Petersen tackled the job of developing microcomputer-based
software that could duplicate the acute
computations of EPA’s mainframe-based
DRES. The prototype program, named “Exposure
4TM”, calculated acute exposure assessments
with means and exposure distributions at
various benchmark percentiles (e.g., 95th,
99th, 99.9th) that were essentially identical
to the acute calculations of DRES for any
defined subpopulation (based on demographic
data in the 1977-78 CSFII). This program
was copyrighted by stephen Petersen in 1988
and licensed to Technical Assessment Systems
for their use and for sublicensing to the
same users who had previously licensed Exposure
1. The prototype acute program required
almost an hour to run on the computer used
to develop it. Today this same program runs
in a few seconds on a late-model, high-end
PC.
Throughout the early 1990’s,
many innovations were added to these Exposure
1 TM and Exposure 4 TM programs, based on
suggestions by users at Technical Assessment
Systems, EPA, the California EPA, and various
private companies that had sublicensed the
software. One of the first enhancements
to the program was the ability to perform
Critical Commodity Contribution Analyses
and Complete Commodity Contribution Analyses
in the chronic mode, which made it much
easier to determine the most important sources
of exposure by RAC and FF. “Plot files”
were generated in the acute mode which allowed
the user to import the arrays of binned
exposure computations in order to improve
the ability to QA the program and to allow
post-processing of these exposure results.
A rudimentary ability to handle Monte Carlo
analysis techniques in acute assessments,
using distributions of pesticide residues
for each RAC/FF rather than deterministic
residue amounts and running multiple iterations
for each individual on each day in order
to achieve stability in the assessment,
provided a more realistic assessment of
the distribution of pesticide exposures
in the subpopulations of interest.
The most challenging
innovation during this period was the migration
from dietary intake data based on the 1977
CSFII to intake data from the 1989-1991
CSFII. This required a team of scientists
at Technical Assessment Systems (TAS) to
create “recipes,” or translation
factors, needed to convert the foods-as-eaten
as reported in the CSFII to their RAC and
FF constituents. While the basis for doing
this had been developed by Dr. Barbara Petersen
at RTI in the early 1980’s, a significant
number of new foods had been introduced
in this new CSFII and the translation process
needed to be improved to incorporate extensive
experience with the older system, requiring
extensive work and QA by the staff of TAS.
stephen Petersen applied the new “recipes”
to the foods-as-eaten for each individual
in the 1989-91 CSFII and created new databases
of both daily intake amounts for each individual
and mean (per capita) intake amounts for
specific subpopulations that could be used
with Exposure programs. Starting in the
mid 1990’s, after a significant amount
of QA, EPA began using the Exposure Series
programs for its tolerance setting activities
for chemical residues in foods.
