The Calendex Program
for Aggregate and Cumulative Exposure Analysis
The passage of the Food Quality
Protection Act (FQPA) in 1996 provided a
new challenge to the pest control industry
and the scientific discipline of risk assessment.
New tools were needed to conduct the analyses
that are now required by FQPA, including
aggregate and cumulative exposure assessments.
Aggregate and cumulative exposure assessments
will result in “worst case”
estimates that are extreme if they are conducted
using default assumptions and simplistic
methodologies. Simplified approaches do
not permit discrimination between risks
that are likely and those that are hypothetical.
An aggregate exposure assessment combines
exposures to pesticide residues on foods
resulting from agricultural use with exposure
to the pesticide resulting from non-agricultural
(i.e., residential) uses of the pesticide
(e.g., termite treatment, turf treatment,
etc.).
In the mid 1990’s Novigen
Sciences scientists Dr. Leila Barraj and
Dr. Marcie Francis developed a “calendar-based”
analytical method for assessing human exposure
to a chemical with multiple uses in a residential
context. This model provided the first probability-based
procedure for evaluating multiple exposures
that avoided the “worst-case”
assumption that an individual would be exposed
to residues from all of the applications
of a chemical at the same time and at maximum
dose levels. This model was for a single
chemical, with fixed application intervals,
and did not include dietary intake. Durango
Software studied this model and brought
over some of the components of the DEEM
dietary exposure model and the CSFII survey
to provide an entirely new program which
could perform dietary and residential exposure
analysis in a “cumulative and aggregate”
context. This new calendar-based program
was called Calendex, and after extensive
QA and user feedback from Novigen Sciences,
it was first made available for sublicensing
to government agencies and end-user institutions
in 1998.
An EPA FIFRA Scientific Advisory
Panel reviewed the Calendex program in 2000
and approved it for use by EPA. Calendex
was the first exposure assessment program
that integrated dietary and residential
(non-dietary) pesticide applications. It
continued to use the 1994-96,1998 CSFII
survey database (providing both demographic
information for a large number of survey
participants and dietary intake records)
and incorporated the use of DEEM residue
files for the dietary analysis. It also
incorporated user-supplied data on the frequency,
amount, and probability of a wide range
of residential pesticide applications as
well as potential human contact amounts
with these pesticides in each application.
From these data Calendex created a calendar
of up to 365 days of duration over which
a wide range of pesticide application, degradation,
and human contact scenarios could be simulated,
route- and chemical-specific exposures computed,
combined exposure amounts estimated day
by day, and resulting exposure distributions
reported for any time period during the
year. In 2001, after the FCID dietary intake
database was incorporated into DEEM-FCID
it was incorporated into a further enhanced
version of Calendex, called Calendex-FCID.
Calendex-FCID was used by EPA to perform
its first aggregate and cumulative exposure
assessments for regulatory purposes (organophosphates)
in 2001-2002.
One of the principal attributes
of Calendex is that it is built on the demographic
and dietary intake database from USDA’s
CSFII, in order to make it consistent with
the dietary analyses conducted over many
years by EPA using the DRES and DEEM programs.
The CSFII provides statistical weighting
factors for each individual who completed
the survey (20,607 in the 1994-96, 1998
CSFII) and with these weighting factors
the entire U.S. population and many subpopulation
can be represented. However, one of the
shortcomings of the 1994-96, 1998 CSFII
is that it only provides two days of food
intake for each individual in the survey.
(Previous CSFIIs only provided 3 days of
intake for each individual so they are not
significantly better.) Since there are no
longitudinal dietary intake databases adequate
for tolerance setting activities, there
are only two practical approaches available
for conducting calendar-based dietary analyses:
(1) use the CSFII intake data for the two
days repeatedly over the year (or at a minimum,
during the corresponding season of the analysis),
selecting new residues for each RAC/FF on
each day; or (2) simulate longitudinal data
using some type of demographic matching
procedure and setting up “cohorts”
of similar individuals to share dietary
intake records throughout the year (adjusting
for body weight differences).
Originally Calendex was based
only on the first approach, as the two days
of CSFII dietary intake data were the best
available validated data available for a
large sample of individuals with identified
demographic characteristics. However, at
the urging of EPA and the SAP reviewers,
the second approach has now been developed
and included with Calendex in version 3.2.
(Both methods can now be used in Calendex
assessments, providing a convenient way
of comparing these two approaches holding
all other variables constant.) Calendex
now can use a “demographic matching
file” of CSFII individuals, where
dietary intake records are shared among
individuals in the same demographic cohort,
and are allocated based on seasonal and
day-of-week information from the original
survey records. A new program called DMFgen,
developed by Durango Software, allows the
user to define the demographic cohorts that
are to be used for sharing dietary intake
records and sets up a “demographic
matching file” with 365 days of dietary
intake records for each of the 20,607 individuals
that completed the 1994-96, 1998 CSFII.
DMFgen uses the USDA statistical weighting
factors from the CSFII in setting up the
shared intake records (thus providing consistency
with the dietary exposure methods used by
EPA in the past). DMFgen provides complete
documentation of the shared intake records
for all individuals, both with regards to
the source of the record and the RAC/FF
amounts, by day, in those records.
While neither of these two
longitudinal approaches (two days sampled
repeatedly through the year, or cohorts
of individuals sharing dietary records)
is entirely satisfactory, they can be used
together to provide alternative exposure
assessments that, along with other risk
assessment methods, can be used to determine
a meaningful range of exposures estimates
for establishing the safety of specific
pesticides and other chemicals in the home
environment and the identification of specific
chemical uses (in an aggregate and cumulative
context) that may be problematic to the
risk assessor.
In addition to the incorporation
of DFM capabilities, other new analytical
capabilities are being added to Calendex
at frequent intervals in an effort to make
this the most useful software for aggregate
and cumulative analysis available to the
risk assessor today and tomorrow. For example,
Calendex can now identify the highest daily
exposure (or highest x-day exposure, where
x can be any number of days up to 91) in
any time period (up to 365 days) for each
individual in the analysis and produce a
number of reports related to this information,
include temporal analysis (the distribution
of these maximum events through the year)
and a critical exposure contribution report
based on these maximum exposure events.
Calendex can now generate maximum and average
quarterly exposure amounts and report means
and percentile distributions of these exposures.
Calendex can now generate a database file
of daily exposure calculations for each
individual in the population of interest
over any number of days during the year
for post processing (using a database program,
for instance) when the risk assessor needs
more information than is can be derived
from standard Calendex reports.
