Monte Carlo Risk Assessment, July 2007

RIVM and Wacheningen University have developed a monte carlo model for estimating risk from dietary intake.This page briefly summarizes what the model is about.

In a Monte Carlo dietary risk assessment the risk of exposure to pesticides or other chemicals from the diet is quantified by combining database information on food consumption with database information from monitoring programs for residues of chemicals in food.

MCRA is intended for users who want to analyze their own consumption and chemical concentration data. MCRA provides the following options
# acute (short-term) risk assessment
# chronic (long-term) risk assessment
# empirical or parametric modelling of residue levels
# modelling of processing effects, unit variability and nondetects levels
# bootstrapping to assess the uncertainty of percentiles
# comparison with deterministic point estimates (IESTI)

For questions please contact:
MailJacob van Klaveren (RIKILT/KAP, residue and food consumption data sets, applications), or
MailHilko van der Voet (Biometris, statistical methods, program development)

http://mcra.rikilt.wur.nl/mcra/mcra.html

Summary

This summary contains information adopted from manuals and reference guides available from the website:  http://mcra.rikilt.wur.nl/mcra/olddocumentation.html

The project is about describing a stochastic (or Monte Carlo) model for dietary risk assessment of chemical
compounds based on monitoring data concerning the quality of foods and agricultural products. Intake
(exposure) assessment is an important step in risk assessment of chemical compounds, such as
agricultural chemicals (pesticides, veterinary drugs), toxins (e.g. mycotoxins) and environmental
contaminants (e.g. dioxins).

MCRA provides the following options:
• acute risk assessment
• chronic risk assessment
• parametric or non-parametric modelling of residue levels
• modelling of processing effects
• modelling of sample variability
• modelling of non-detects levels
• restrictions on age and/or days
• consumers only

MCRA is available as standalone version or as internet application.

The program MCRA is composed of a set of procedures which may be arranged into four
main blocks. The main tasks of block 1 to 4 are:
1. reading of data (residue concentration data, consumption data, consumer characteristics,
processing factors, variability factors, percent crop treatment)
2. pre-processing of datastructures (age and/or day restrictions, consumers only, processing or not,
variability factors, estimation of parameters for a parametric model), determining number of
loops, chunksize, etc.
3. simulation of exposure values (parametric, non-parametric)
4. generating output (intake distribution, contribution to upper tail, characteristics of consumers with
the highest intake, etc…)

MCRA is a computational tool for dietary risk assessment. MCRA can
calculate intake distributions for both short-term (acute) and long-term (chronic) intakes. Basically, it
simulates daily consumptions by sampling a food consumption database and combines these with a
random sample from either a compound database (empirical distribution) or a parametric distribution
of compound concentrations. The result is a full distribution of intakes, rather than traditional
deterministic methods which only provide a point estimate. Percentiles of the intake distribution can
be used to assess risks by relating them to e.g. an acute reference dose (ARfD). In a chronic risk
assessment, MCRA calculates the distribution of the usual intakes over consumers based on the
average concentration and the empirical distribution of intake between consumers and between
different intake days of the same consumers. Percentiles of this usual intake distribution can then be
related to e.g. the acceptable daily intake (ADI). Uncertainty of percentiles can be established by
resampling methods. MCRA allows including processing factors (e.g. the effect of cooking on the
concentration) and variability factors (to correct for the fact that monitoring data are obtained from
composite samples, whereas consumers may eat individual units). Analyses can be done for a total
population or for a subpopulation (e.g. children, males or females or consumption-days only). The
effects of concentration below analytical reporting limits (LOR) can be assessed. Large portion
consumption and the highest compound or median compound in case of bulking or blending in the
composite sample is used in IESTI (International Estimated Short Term Intake) calculations.